CHANGELOG.md

CHANGELOG

Inspired from Keep a Changelog

Unreleased

Added

• Enrich opensearchtransport.RequestEvent with request identity for observers: RouteName (the classified operation, e.g. "search"/"doc_index", suitable as a low-cardinality metric label; "other" when unclassified), Index (target index extracted from the path; now populated where it was previously documented but always empty), and PoolName (the pool that served the final attempt). Path becomes the raw URL-escaped caller input (req.URL.EscapedPath()) captured before the URL is rewritten to the selected backend, with a corrected doc comment (it is raw and high-cardinality, not templated). Identity is captured once per request, pre-rewrite, and only when an observer is registered, so the no-observer path stays allocation-free. RouteName is derived from a process-wide OperationClassifier; callers routing non-standard paths can override it via the new Config.OperationClassifier (affects the label only, never routing). Route the previously-unrouted system/admin endpoints (cluster state, _cat/*, _nodes/*, _tasks/*, _snapshot/*, index/mapping/alias/template management, data streams) to their correct node role, adding a cluster-manager routing target (split read/write pools) so cluster-state operations prefer cluster_manager nodes with a data-node fallback; because the router and the operation classifier share the route table, these routes also give each operation a meaningful RouteName. Name every OperationID.String() label as a constant, and add explicit labels for the admin operations (cat/nodes/tasks/snapshot/script/dangling/data-stream) that previously fell through to the numeric admin_N fallback. Rework the osprom and osotel registries around an event-sink Observer (two-method OnRequestResponse/OnStreamResponse plus connection-lifecycle hooks, receiving the full transport event so sinks can label by route/index/pool; RequestSample is removed), dispatched by a pool of workers (GOMAXPROCS/2 by default, WithWorkers to override). Ship RED+USE defaults out of the box: RequestObserver (rate, errors, duration -- with a WithStatusClassifier option to control the status label; the default maps to the per-hundred bucket 2xx/3xx/4xx/5xx, error for no response, or unknown) and PoolObserver (connection-pool utilization, saturation, and errors from lifecycle events). Add WithRequestFilter/WithStreamFilter (skip events before enqueue) and WithOverflowHandler/WithStreamOverflowHandler (invoked on buffer overflow with the queue length -- at most the buffer size -- and the dropped event, in addition to the existing dropped counter) (#999)
• cmd/osgen: emit enum-like oneOf-of-{type: string, const: X} schemas as named Go string types with one exported const per value (e.g. type NodeRole string with NodeRoleData, NodeRoleSearch, NodeRoleWarm, ...), instead of collapsing them to a plain string. The type is permissive: backed by string, encoding/json round-trips any value, so an unknown value a newer server or plugin introduces decodes cleanly rather than erroring (unlike the closed int-backed RestStatus enum, whose UnmarshalJSON rejects unknown values). No custom (Un)MarshalJSON is generated; the consts add discoverability, compile-time typo protection, and spec-driven drift detection. Detection is default-on for the const-oneOf shape (deny-list to opt out); branches are version-filtered, then reduced to values yielding a valid, unique Go const identifier and deduplicated by generated const name (collapsing NodeRole's duplicate search across a version boundary and TranslogDurability's ASYNC/async casing collision, reporting a dropped distinct value to stderr). Query parameters whose schema is a const-oneOf are typed too. Also bundles the search and warm node roles into the vendored spec (#998)
• Export the opensearchapi and plugin sub-client types (documentClient -> DocumentClient, catClient -> CatClient, plugin roleClient -> RoleClient, etc.), so each sub-client renders its own godoc page with a navigable method list instead of being an unexported type reachable only through a Client field. The change is additive and non-breaking: field access (client.Doc.Get) is unchanged, the apiClient/client back-pointer stays unexported so external populated construction remains impossible, and each generated type carries a doc comment noting it must be obtained via NewClient (the zero value is not usable). Adds a receiver-to-sub-client map to the opensearchapi package doc and generated package docs to the plugin packages that have sub-clients (#989)
• Add VerifyDeadAfter (opensearch.Config / opensearchtransport.Config, env override OPENSEARCH_GO_VERIFY_DEAD_AFTER): bounds how long a connection proven reachable may still be blindly resurrected as a last-resort "zombie" while dead. Each discovery cycle clears the viability mark on any non-seed connection that has been dead longer than the window, so a node that never recovers stops absorbing requests and must health-check clean again before it is routed to; seed connections are exempt. The env var accepts a boolean (true selects the 15m default, false disables the expiry) or a time.ParseDuration string; the Config field follows the 0 = default, <0 = disabled, >0 = explicit convention. See guides/transport-routing.md (Zombie Connection Resurrection and Connection Viability) and guides/config-envvars.md
• Add response observer events to opensearchtransport.ConnectionObserver: OnRequestResponse(RequestResponseEvent) fired once per request by Transport.Request (full-read duration, exact response-byte count) and OnStreamResponse(StreamResponseEvent) fired by Transport.Stream (time-to-first-byte, Content-Length header). Both events are flat value types passed by value, so the fire path is allocation-free and a nil observer costs nothing; BaseConnectionObserver gains no-op defaults so existing observers are unaffected. Adds opensearchtransport.Transport.Request (and an opensearch.Client.Request passthrough) as the buffered counterpart to Stream, through which opensearch.Execute[T] now routes
• Add the osprom module (a separate Go module, so github.com/prometheus/client_golang stays out of the core dependency graph): an osprom.Registry is the single ConnectionObserver wired into a transport, into which callers wire a Prometheus registerer and an arbitrary set of Observer bundles. The Registry copies each event into a pooled envelope and fans it out to the wired observers on a background goroutine (New + Run(ctx) + Close), so recording metrics never blocks or allocates on the request hot path; a full buffer drops events and increments opensearch_client_observer_dropped_total rather than adding latency. Ships NewRequestObserver, a bundle recording request-duration and response-size histograms labeled by method, status class, and mode. See osprom/README.md and guides/transport-observer_metrics.md
• Add the osotel module, the OpenTelemetry counterpart to osprom with the same Registry + Observer model and async pooled pipeline (a separate Go module, so the OpenTelemetry libraries stay out of the core dependency graph). Callers wire an OpenTelemetry metric.Meter and an arbitrary set of Observer bundles into osotel.New; the shipped NewRequestObserver records opensearch.client.request.duration and opensearch.client.response.size histograms attributed by method, status class, and mode, and the Registry records opensearch.client.observer.dropped. See osotel/README.md
• opensearchtransport RouteEvent now reuses a pooled []RouteCandidate backing array across routing decisions, reference-counted so an observer that retains the event past OnRoute (for example, for async processing) calls event.Retain() then event.Release() when done; synchronous observers are unaffected and the dispatch path allocates zero candidates per route
• cmd/osapilint: add the v2 -> v3 hop, so a v2 module can be migrated through to v5. The v2 -> v3 boundary is the largest in the project's history and mostly not mechanizable: the opensearchapi package was redesigned from a function-based API into a typed sub-client API, so of 182 exported v2 structs only 16 survive by name and 166 (the entire opensearchapi.*Request family) are removed outright. The hop bumps import paths (the only mechanical change), rules the 51 removed root-opensearch.Client API method fields MANUAL with actionable migration guidance, and relies on the linter's removed-type diagnostic to surface any reference to a removed opensearchapi.*Request type. It best-effort rewrites the two seed ops Ping and Indices.Exists (the call, its raw-response Status()/Warnings handling, and the Config/NewClient client lifecycle) into compiling v3; where a transform is not mechanically certain it plants an _OSAPILINT_RESOLVE marker so the build breaks at that spot instead of emitting a wrong value. The two consumer idioms and the raw-Response-to-typed-Resp error-model change are reported as semantic followups. The removed-type diagnostic runs on every hop rather than only v2 -> v3, so a reference to a type deleted across any transition (for example the many opensearchapi types dropped in v5) is reported as a MANUAL worklist item instead of being silently dropped; it fires only when a consumer actually references the removed type. See cmd/osapilint/README.md (#951)
• Add cmd/osapilint, a tool that migrates a Go module across opensearch-go major versions. osapilint rewrite -w ./... detects the source major from the module's imports and rewrites old-major API shapes (type renames, field dispositions, method regroups) into the target's; the pass is purely syntactic (go/parser + astutil + go/printer), so it runs before the code compiles against the target, and writes are sandboxed to the target module directory via os.Root. osapilint vet -fix ./... then runs go/analysis analyzers that catch runtime type-assertion hazards -- the target's precise *int64/*string types flowing into any sinks such as testify's Equal, which compile cleanly but panic at run time. Each adjacent version transition is a hand-authored Hop keyed against two committed API surfaces (surface_vN.json); a migration resolves to the ordered list of hops between source and target, and new hops (v2 -> v3) are added as data without linter changes. Ships the v3 -> v4 and v4 -> v5 hops, chainable to migrate v3 -> v5 directly. See cmd/osapilint/README.md (#933)
• Add Close() to opensearch.Client and opensearchapi.Client for explicit teardown of background goroutines (node discovery, health/stats pollers, DNS refresh) and idle connections, without type-asserting the transport. Cache implicitly-constructed default clients (opensearch.NewDefaultClient, opensearchapi.NewDefaultClient, and the client opensearchutil.NewBulkIndexer builds when none is supplied) in a process-wide, refcounted, idle-TTL cache keyed by config hash, so identical default clients share one transport instead of leaking one set of goroutines and its connection pool per construction. User-built opensearch.NewClient/opensearchapi.NewClient clients never enter the cache. opensearchutil.NewBulkIndexer now closes the client it implicitly creates when the indexer is closed. Tune the idle eviction window with OPENSEARCH_GO_DEFAULT_CLIENT_TTL (default 16m; 0 = never evict; a negative value disables caching so every call builds a fresh client) (#893)
• Add client-side DNS caching, enabled by default on the built-in transport. Resolved addresses are cached and re-resolved on an interval (default 60s, mirroring the TTL AWS publishes for managed OpenSearch Service endpoints). When the resolver becomes briefly unreachable, the last-known-good address continues to be served until the resolver recovers, so transient resolver outages (e.g. a node-local DNS blip producing dial tcp: lookup ...: i/o timeout) no longer fail requests for already-resolved hosts. Tune or disable via the DNSCacheRefresh, DNSDialTimeout, DNSKeepAlive, and DNSTimeout fields on opensearch.Config (or OPENSEARCH_GO_DNS_CACHE_REFRESH, OPENSEARCH_GO_DNS_DIAL_TIMEOUT, OPENSEARCH_GO_DNS_KEEP_ALIVE, OPENSEARCH_GO_DNS_TIMEOUT); each follows the 0 = default, <0 = disable, >0 = explicit convention. Caching is installed only when no custom Transport is supplied; a caller-provided Transport is never modified. A host that resolves to multiple addresses races up to three of them concurrently (random start offset per connection) and takes the first to connect, spreading load and tolerating a dead address. Refresh re-resolves cached hosts sequentially, so DNSTimeout (default 10s) bounds each lookup to keep one hung resolution from stalling a refresh tick. The refresh goroutine is bound to the client's root context, so it is reclaimed both when Close is called and when New returns an error after the context is created. Because Go's resolver does not expose record TTLs, the refresh interval is a re-resolution cadence, not a per-record TTL. Exposes DNSLookups, DNSCacheMisses, and DNSLookupErrors counters via Transport.Metrics()
• cmd/osgen: guard json.RawMessage in generated request/response types behind a checked-in allowlist (cmd/osgen/rawmessage_allowlist.txt). Because a json.RawMessage is the symptom of a type the generator could not resolve, a generator bug can silently widen the raw-JSON surface of the public API; generation now fails (non-zero exit) when any json.RawMessage use is not listed, including nested forms such as []json.RawMessage, map[string]json.RawMessage, and [][]json.RawMessage (the leaf is detected at any wrapper depth). Entries are keyed GoTypeName/jsonFieldName (whole-response raw bodies use <Prefix>Resp/-, and map/array responses whose element type is unresolved use <Prefix>Resp/[entries] and <Prefix>Resp/[records]). Add -update-raw-message-allowlist to regenerate the allowlist from current output (sorted and grouped for minimal diffs), and -allow-unlisted-raw-message to downgrade the check to a warning (#890)
• cmd/osgen: emit int-backed (const iota) enum types for string fields carrying an x-enum-name marker alongside an enum: constraint. Each enum generates a named int type with a zero-value <Name>Unknown sentinel, name<->value lookup maps, String(), MarshalJSON, and a closed-set UnmarshalJSON that rejects unknown wire values via a typed *Unknown<Name>Error (recoverable through errors.As). The marker is shared, so a single enum type is registered once and reused across every referencing field; a marker reused with a conflicting value set fails generation rather than silently merging. Applied to the security status field, which becomes a typed RestStatus enum (#890)
• Add OPENSEARCH_GO_POLICY_DUMP environment variable: when set with OPENSEARCH_GO_DEBUG=true, dumps the router's policy tree (the dot-delimited node paths that OPENSEARCH_GO_POLICY_* matchers target, each labeled with its pool or role) to the debug logger at client initialization. The dump walks the structural tree so router wrappers that share an inner policy instance are each rendered in full. (#883)
• Add a build-samples Makefile target and a CI job that compiles and vets every _samples/*.go program, so example breakage is caught (the _samples directory is excluded from go build ./... because Go ignores _-prefixed paths)
• Group document operations under a client.Doc sub-client and point-in-time operations under client.PIT (Create/Delete/GetAll/DeleteAll); client.Document and client.PointInTime remain as field aliases. The indices sub-client's canonical field is client.Index, with client.Indices and client.Indexes as aliases. cmd/osgen gains --emit-v4-compat (default true) to emit backward-compatibility forwarders so top-level client.Bulk/MGet/Update, client.Document.Source, and client.PointInTime.Get keep working (client.Index is not forwarded -- it is the indices sub-client field; use client.Doc.Index), and --emit-v4-deprecation (default false) to mark those forwarders deprecated
• Add cmd/osgen code generator for typed path builders and API consumer files from the OpenAPI spec
• The built-in transport builds opensearchtransport.NewDefaultRouter when Config.Router is nil, and opensearch.NewClient enables on-start discovery under the same condition, so a client routes across discovered nodes by default. OPENSEARCH_GO_ROUTER controls it: =false/=0 suppresses both the default router and on-start discovery; unset or any other value builds the router and enables discovery. (#816)
• Add envvars.Falsy(name), which tells "explicitly opted out" apart from "unset" (Truthy treats both as false). The default-router rule in the transport and opensearch.NewClient use it.
• Add the code-generated opensearchapi/ package: API surface produced by cmd/osgen from the OpenAPI spec. Fully typed Req/Resp/Params structs, sub-clients matching OpenSearch namespaces (client.Cat, client.Cluster, client.Indices, etc.), and a plugins/ subtree for ML/k-NN/security/ISM/etc. Replaces the hand-written v4 package (previewed in the v4 line at v5preview/opensearchapi/); see opensearchapi/README.md for usage and UPGRADING.md for migration guidance (#650)
• Add primary_terms_map and split_shards_metadata fields to ClusterState index metadata for OpenSearch >=3.6.0 compatibility
• Add address resolver handler to rewrite discovered node addresses before they enter the connection pool (#822)
• Add InsecureSkipVerify config option to disable TLS certificate verification without constructing a custom http.Transport, preserving DefaultTransport connection pooling, HTTP/2, and timeout defaults (#786)
• Add (*opensearchtransport.Transport).Stream(*http.Request) (*http.Response, error) and a (*opensearch.Client).Stream passthrough for raw byte forwarding (incremental/streaming use cases). Stream returns the unbuffered response body from RoundTrip; the caller owns reading and closing res.Body. Pairs with opensearch.Execute[T] for typed, decoded results (the SDK owns the body) (#786)
• Add per-attempt RequestTimeout to bound individual HTTP round-trips, preventing indefinite hangs on stalled connections (#786)
• Add opensearchutil/shardhash package with exported Hash and ForRouting functions for computing OpenSearch shard routing
• Enhanced cluster readiness checking for improved test reliability: testutil.NewClient() now includes readiness validation (health + cluster state + nodes info)
• Add Status field (json.RawMessage) to TasksGetResp, TasksListTask, and TaskCancelInfo for polymorphic task status data; add typed status structs matching the OpenSearch API specification: BulkByScrollTaskStatus, ReplicationTaskStatus, ResyncTaskStatus, PersistentTaskStatus; add Parse* helpers and BulkByScrollTaskStatusOrException for sliced task status (#788)
• Test parallelization support via TEST_PARALLEL environment variable (default: CPU cores - 1, minimum 1)
• Add cmd/osgen/emit.TestPerOpErrorTypeName_CatalogConsistency to pin the catalog <-> switch coupling between emit.PerOpErrorTypeName and errwrap.OperationWrappers. Asserts three directions: every group naming a per-op aggregator type has 2+ wrappers in the catalog, every catalog entry with 2+ wrappers names a per-op aggregator type, and every group named by the switch is present in the catalog. Does not pin the runtime emittableWrappers/resolveErrorWrappers paths; today those sets coincide for the only 2+-wrapper groups (msearch / msearch_template) (#857)
• opensearchapi/testutil package with test suite, client helpers, and JSON comparison utilities
• Add typed path builders in internal/path/ generated from the OpenAPI spec via cmd/osgen for compile-time URL construction safety (#617, #650)
  • sync.Pool-backed []byte buffers eliminate per-request allocation churn; buffers over 4 KiB are discarded to bound pool growth
• opensearchtransport/testutil package with PollUntil helper for eventual consistency testing (ISM policies, index readiness, cluster state changes)
• Configuration option IncludeDedicatedClusterManagers for controlling cluster manager node routing (#765)
• Policy-based routing system for improved request routing and service availability (#771)
  • Policy interface for composable routing strategies with lifecycle management
  • Router interface with Route() method for request-based connection selection
  • NewPolicy() implementing chain-of-responsibility pattern for composable routing strategies
  • NewIfEnabledPolicy() for conditional routing with runtime evaluation
  • NewMuxPolicy() for trie-based HTTP pattern matching with zero-allocation route lookup
  • NewRolePolicy() for role-based node selection
  • NewRoundRobinRouter() with coordinating node preference and round-robin fallback
  • NewMuxRouter() providing role-based request routing with graceful fallback
    • Automatic routing of bulk, streaming bulk, and reindex operations to ingest nodes
    • Automatic routing of search operations (search, msearch, count, by-query operations, scroll, PIT, validate, rank eval) to search/data nodes
    • Automatic routing of document retrieval operations (get, mget, source, explain, termvectors, mtermvectors) to search/data nodes for read locality
    • Automatic routing of template operations (search template, msearch template) and search shards to search/data nodes
    • Automatic routing of field capabilities to search/data nodes
    • Automatic routing of shard maintenance operations (refresh, flush, synced flush, forcemerge, cache clear, segments) to data nodes
    • Automatic routing of single-document writes (index, create, update, delete) to data nodes
    • Automatic routing of shard diagnostics (recovery, shard stores, stats) and rethrottle operations to data nodes
  • NewDefaultRouter() extending role-based routing with per-index node affinity (recommended for most users)
• Add consistent hash routing with per-index node affinity for cache locality and AZ-aware load distribution (#786)
  • Rendezvous hashing selects a stable subset of nodes per index, preserving OS page cache and query cache locality
  • RTT-bucketed scoring naturally prefers AZ-local nodes and overflows to remote AZs under load
  • Per-pool congestion window (cwnd) routing using TCP-style AIMD congestion control for capacity-aware connection scoring
  • Thread pool discovery via /_nodes/_local/http,os,thread_pool provides per-pool capacity ceiling (maxCwnd)
  • Thread pool stats polling via /_nodes/_local/stats/jvm,breaker,thread_pool drives AIMD window adjustments
  • Scoring formula: RTTBucket * (InFlight + 1) / Cwnd * ShardCostMultiplier -- nearest node with most headroom wins
  • In-flight request tracking per node per pool with atomic add/release bracketing each RoundTrip
  • AIMD slow start (double cwnd) transitions to congestion avoidance (additive increase) at ssthresh
  • Multiplicative decrease on congestion signals: total_wait_time_in_nanos for RESIZABLE pools, queue saturation fallback for others
  • Pool overload detection: delta(rejected) > 0 or HTTP 429 marks pool overloaded, cleared only by stats poller
  • HTTP 429 handling: TryLock + set overloaded + halve cwnd + retry on different node
  • Quorum gating: pre-quorum uses 4 * defaultServerCoreCount (= 32) as synthetic cwnd, post-discovery uses 4 * allocatedProcessors, post-quorum uses real pool cwnd
  • Asymmetric scale-up/scale-down thresholds with hysteresis band for stable active pool sizing
  • Dynamic per-index fan-out driven by shard placement (/_cat/shards) and request rate
  • RouterOption functional options: WithMinFanOut, WithMaxFanOut, WithIndexFanOut, WithIdleEvictionTTL, WithDecayFactor, WithFanOutPerRequest
• Add environment variable escape hatches (OPENSEARCH_GO_POLICY_*) to disable specific routing policies at startup (#786)
• Add failure-triggered shard map invalidation for faster routing recovery (#786)
  • lcNeedsCatUpdate lifecycle bit excludes failed connections from routing candidate sets until /_cat/shards refresh
  • Connections remain available for general routing (round-robin, zombie tryouts) while excluded from scored routing
  • Dedicated discoverCatTimer schedules lightweight /_cat/shards-only refresh (no full node discovery)
  • Refresh urgency scales with cluster impact: interval = discoverNodesInterval * (1 - flaggedFraction), clamped to 5s floor
  • OnShardMapInvalidation observer event for monitoring invalidation triggers
  • NeedsCatUpdate field in ConnectionMetric for observability
• Add routing observability: observer events, metrics snapshot, connection inspection (#786)
  • OnRoute observer event with full scoring breakdown (RouteEvent, RouteCandidate)
  • RouterSnapshot in Client.Metrics() exposes per-index cache state (fan-out, shard nodes, request rate, idle-since)
  • Connection.RTTMedian(), Connection.RTTBucket(), Connection.EstLoad() for per-connection inspection
  • ConnectionMetric enriched with rtt_bucket, rtt_median, est_load fields
• Add murmur3 shard-exact routing for ?routing= and document ID requests (#786)
  • Client-side murmur3 x86 32-bit hash matching OpenSearch's Murmur3HashFunction.hash(String) for shard-exact targeting
  • Document-level requests (_doc, _source, _update, _explain, _termvectors) use doc ID as default routing value when no explicit ?routing= present, matching OpenSearch's OperationRouting.generateShardId() behavior
  • Client-side murmur3 shard-exact candidate selection routes requests to nodes hosting the target shard
  • Per-shard-number placement data from /_cat/shards maps shard numbers to primary and replica node names
  • Graceful fallback to rendezvous hashing when shard map data is unavailable
  • RoutingValue, EffectiveRoutingKey, TargetShard, ShardExactMatch fields in RouteEvent for observability
• Add OPENSEARCH_GO_ROUTING_CONFIG and OPENSEARCH_GO_DISCOVERY_CONFIG environment variables for runtime feature control (#786)
  • OPENSEARCH_GO_ROUTING_CONFIG: toggle shard-exact routing (-shard_exact)
  • OPENSEARCH_GO_DISCOVERY_CONFIG: skip individual discovery server calls (-cat_shards, -routing_num_shards, -cluster_health, -node_stats)
  • Bitfield flags use +/- prefix convention for explicit opt-in/out; zero-initialized = all features enabled
  • WithShardExactRouting(bool) RouterOption for programmatic control (env var overrides)
  • Evaluated once at client init time; immutable after
  • Document environment variables in guides/transport-routing.md
  • Document read-after-write visibility guarantees with operation-aware routing in guides/transport-routing.md
• Add adaptive max_concurrent_shard_requests derived from cluster-wide AIMD congestion window (#800)
• Add partial failure error types (PartialBulkError, PartialSearchError, ShardFailureError, MultiSearchItemError) that surface HTTP 200 partial failures as typed Go errors, controlled by a per-category errmask.ErrorMask bitfield on Config.Errors (#816)
  • PartialBulkError returned from Bulk when resp.Errors is true, carries FailedItems and SucceededCount
  • PartialSearchError returned from Search, MSearch, MSearchTemplate, SearchTemplate, Scroll.Get when _shards.failed > 0
  • ShardFailureError returned from Index, Document.Create, Document.Delete, Update when replica shards fail
  • MultiSearchItemError returned from MSearch/MSearchTemplate for per-sub-response Error envelopes
  • MSearchErrors / MSearchTemplateErrors per-op containers (Go 1.20+ multi-error contract via Unwrap() []error) when 2+ wrapper categories fire on the same response
  • PartialFailureError marker interface with IsPartial() bool for type-switching across all partial-failure types
  • opensearchapi.Errors(err) []error package-level helper that flattens single- and multi-wrapper errors into a uniform slice; recommended call-site pattern is a for/switch over the result (not errors.As against a specific type)
  • Helper functions: IsPartialFailure, ToleratePartialFailures, RequireSuccessRate for threshold-based error tolerance
  • Operation constants: OperationIndex, OperationCreate, OperationUpdate, OperationDelete
  • Per-Resp helper methods (BulkItemFailures, SearchShardFailures, WriteShardFailures, MultiSearchItemFailures, PartialFailures(mask)) exist on the response types as engine machinery for the dispatch; new code should prefer a for/switch over opensearchapi.Errors(err) rather than the per-Resp helpers, for forward compatibility
  • Config.Errors *errmask.ErrorMask replaces a single boolean: each bit suppresses one wrapper category. v4 defaults to errmask.All (mask everything, preserves pre-bitfield behavior); v5+ defaults to errmask.Empty (report everything)
  • OPENSEARCH_GO_ERROR_MASK environment variable overrides Config.Errors at runtime via comma-separated +/- tokens (lowercase snake_case wrapper names; unknown tokens silently dropped, debug-logged)
  • Both (resp, error) are non-nil on partial failure -- response is fully populated
  • The generated opensearchapi uses spec-driven types for the same model (regenerated from the OpenAPI x-error-responses extension on every cmd/osgen run)
• Add OperationClassifier for zero-allocation HTTP method+path to OperationID mapping (#816)
  • Bit-packed OperationID (int64) encoding R/W flag, category, and minor operation
  • Masking helpers: IsWrite, IsRead, Category, Minor
  • String() returns Prometheus-friendly labels (e.g., "search", "bulk", "doc_get")
  • Reuses existing routeTrie for O(path-segments) lookup, safe for concurrent use
  • Enables transparent metrics/tracing middleware at the http.RoundTripper layer
  • Transport automatically sets max_concurrent_shard_requests query parameter on search requests routed through a coordinator node
  • Value derived from a cluster-wide aggregate of all polled nodes' search pool wait-time and completion deltas, clamped to [floor, cap] (default: 5–256)
  • Cluster-wide signal correctly models data-node fan-out capacity: single hot nodes are diluted by healthy peers, and MCSR only drops when aggregate cluster pressure rises
  • Per-node AIMD for connection scoring remains unchanged (hot-node avoidance is handled by connection selection, not fan-out throttling)
  • Falls back to per-node cwnd before the first poll cycle completes
  • Respects explicit caller overrides: pre-existing max_concurrent_shard_requests query parameter is never clobbered
  • Not applied to shard-exact routed requests (coordinator fan-out is irrelevant)
  • WithAdaptiveConcurrency(bool) and WithAdaptiveConcurrencyLimits(floor, cap) RouterOption for programmatic control
  • OPENSEARCH_GO_SHARD_REQUESTS environment variable: true/false to enable/disable, or min:max to set floor and cap (e.g., 10:512)
  • OPENSEARCH_GO_ROUTING_CONFIG=-adaptive_mcsr to disable via routing config bitfield
  • MaxConcurrentShardRequests field in RouteEvent for observability
• Add seed URL fallback as last-resort connection source when all router pools are exhausted (#786)
  • Builds a dedicated multiServerPool from fresh copies of the original seed URLs at client init
  • Fires after the entire retry loop when all router policies and connection pools return ErrNoConnections
  • On success: triggers immediate cluster rediscovery to repopulate router pools
  • OPENSEARCH_GO_FALLBACK=false disables seed fallback (enabled by default)
• Add consolidated environment variable reference in guides/transport-routing.md and USER_GUIDE.md (#786)
• Add connection pool health probes with cluster-aware resurrection timing (#786)
  • Auto-discover server core count from /_nodes/http,os,thread_pool to derive all rate-limiting and congestion window parameters (default: 8 cores)
  • Weighted round-robin for heterogeneous clusters: nodes with more cores get proportionally more traffic via GCD-normalized duplicate pointers in the ready list
  • lcNeedsHardware lifecycle bit tracks connections needing hardware info; per-node fallback via /_nodes/_local/http,os,thread_pool during health checks
  • Capacity model dynamically recalculated on each discovery cycle from minimum allocatedProcessors across all nodes
  • TLS-aware rate limiting prevents overwhelming recovering servers during outages
  • Three-input timeout formula: max(healthTimeout, rateLimitedTimeout, minimumFloor) + jitter
  • Shuffle ready connection list on add/resurrect to prevent round-robin hot-spotting
  • Two-phase readiness health check: GET / then GET /_cluster/health?local=true with initializing_shards gate to prevent routing to recovering nodes
  • Store cluster health metrics (ClusterHealthLocal) on each connection for observability
  • Periodic cluster health refresh for ready connections keeps ClusterHealthLocal data current for load-shedding and routing decisions
    • Refresh interval scales with cluster size: clamp(liveNodes * clientsPerServer / healthCheckRate, 5s, 5min)
    • Single-node clusters skip refresh entirely (no routing benefit)
  • Node stats polling with load shedding via NodeStatsInterval configuration
    • Polls GET /_nodes/_local/stats/jvm,breaker,thread_pool to detect overloaded nodes and update congestion windows
    • Per-pool AIMD congestion control adjusts cwnd based on thread pool wait time and queue saturation
    • Overloaded nodes are demoted from the ready list to the standby partition
    • Overload detection: JVM heap threshold (OverloadedHeapThreshold, default 85%), circuit breaker size ratio (OverloadedBreakerRatio, default 0.90), breaker trip delta, and cluster status red
• Add heterogeneous Docker cluster targets for integration-testing weighted routing and role-based request routing
  • cluster.heterogeneous.cpu.1 and cluster.heterogeneous.cpu.2 set per-node CPU limits via Docker Compose overrides
  • cluster.heterogeneous.roles assigns distinct node roles (cluster_manager+ingest, data+ingest, data)
  • cluster.homogeneous removes all overrides to reset to default configuration
  • cluster.status now shows per-node roles and allocated processors via _nodes/http,os
• Add request routing guide (guides/transport-routing.md) consolidating routing architecture, connection scoring, pool lifecycle, cost model, and configuration reference (#786)
• Add per-item Error field to MGetResp, MTermvectorsResp, and MSearchResp for detecting partial failures in multi-document operations (#797)
• Add DocumentError type for structured per-item error information in multi-document responses
• Add BulkByScrollFailure type for structured failure information in _delete_by_query, _update_by_query, and _reindex responses
• Add Routing and Fields to MGetResp.Docs to match the full OpenSearch _mget response format
• Add ForcedRefresh field to IndexResp, DocumentDeleteResp, and UpdateResp for consistency with DocumentCreateResp
• Add Status and Primary fields to ResponseShardsFailure for shard failure diagnostics
• Add guides/config-envvars.md as the canonical reference for every OPENSEARCH_GO_* environment variable — accepted values, defaults, parsing rules, and the exhaustive OPENSEARCH_GO_ERROR_MASK token list. Fix OPENSEARCH_GO_ROUTER default in transport-routing.md from incorrect false to correct true. (#883)

Changed

• BREAKING: cmd/osgen now types enum-like oneOf-of-const fields and query parameters across opensearchapi and the plugin packages, so fields and params that were string change to named enum types (e.g. Roles []string -> Roles []NodeRole, Result string -> Result Result, the cat time param string -> TimeUnit, and similarly OpType, SearchType, VersionType, ...). Values still assign and compare as strings through the named type, so most call sites are unaffected; only type-strict comparisons against untyped string literals need the const (e.g. require.Equal(t, opensearchapi.ResultUpdated, resp.Result) instead of "updated"). (#998)
• cmd/osapilint: map the v4 -> v5 partial-failure type renames from opensearchapi/UPGRADING_V4_TO_V5.md in the v4 -> v5 hop, so code walking the per-shard failure slice migrates instead of silently breaking. ResponseShards -> ShardStatistics and ResponseShardsFailure -> ShardSearchFailure (dropping the removed Primary/Status fields) are rewritten; DocumentError -> ErrorRespBase is reported as a manual semantic followup because the two types share no fields. (#963)
• Add a first-class container-provider abstraction to the test harness Makefile. CONTAINER_PROVIDER is auto-detected by CLI presence in the order Colima -> Rancher Desktop (rdctl) -> Docker, and overridable with CONTAINER_PROVIDER=colima|rancher|docker. Selecting a provider pins the docker context (colima / rancher-desktop; the Docker provider leaves the active context alone, and a pre-set DOCKER_CONTEXT in the environment is respected), resolves the CLI runtime $(CTR) (now docker by default for every provider, with CONTAINER_RUNTIME=nerdctl as an advanced override), ensures the backing VM/daemon is running via the new cluster.provider.ensure target (wired into cluster.start), and sets vm.max_map_count through the provider's VM (colima ssh / rdctl shell) or a privileged helper container. Previously $(CTR) preferred nerdctl whenever it was on PATH, so a Rancher-installed nerdctl could hijack a Colima session. make cluster.runtime now reports the detected provider, docker context, and runtime
• BREAKING: Per-request transport metrics (requests, failures, responses-by-status) are now always collected via lock-free atomics, independent of EnableMetrics. EnableMetrics now gates only the detailed-metrics snapshot (per-connection, per-policy, and router state returned by Metrics()). The responses-by-status counter moved from a mutex-guarded map to a lock-free atomic array. Metrics() no longer returns an error when metrics are disabled -- it always returns the per-request counters (callers that branched on if err != nil for the disabled case should drop that check). See UPGRADING_V5.md for migration. (#891)
• Make the detailed-metrics snapshot path lock-free at call time. The per-connection deadSince/overloadedAt timestamps moved from Connection.mu-guarded time.Time fields to lock-free atomic Unix-nanosecond values, so Metrics() enumerates connections without taking each connection's mutex. Under concurrent request load this was the dominant lock-contention site (a mutex profile attributed ~3.85% of total contention delay to the snapshot reader taking a write lock merely to read two fields); the conversion drops that to ~0.1%. Writes still occur under Connection.mu so the resurrection/standby read-modify-write decisions stay serialized. Benchmarks (BenchmarkMetrics, BenchmarkMetricsParallel, BenchmarkMetricsUnderLoad) confirm the always-on detailed path is acceptable. (#892)
• BREAKING: Implicitly-created default clients are now cached and shared. Two opensearch.NewDefaultClient (or opensearchapi.NewDefaultClient) calls with identical config resolve to one shared transport instead of two independent ones, so they share goroutine and connection-pool lifecycle and their Metrics() counters are aggregated across all holders rather than per-client. A caller that built multiple default clients to read separate metrics will now see combined counts. To keep independent transports, build with opensearch.NewClient/opensearchapi.NewClient (never cached) or set OPENSEARCH_GO_DEFAULT_CLIENT_TTL to a negative value to disable caching. See UPGRADING_V5.md for migration. (#893)
• Reorganize the documentation. Split UPGRADING.md into a version-history index plus per-major-version guides (UPGRADING_V5.md through UPGRADING_V2.md) and rename opensearchapi/MIGRATING.md to opensearchapi/UPGRADING_V4_TO_V5.md. Group the guides/ and _samples/ files by subsystem (transport-, indexing-, usage-, config-) and add a guides/README.md index. Make guides/usage-error_handling.md the single source for partial-error handling and guides/transport-retry_backoff.md the single source for resurrection-timeout config, replacing the duplicated copies in opensearchapi/README.md and guides/transport-routing.md with links. Add package documentation (doc.go) for opensearchapi, plugins, signer, and signer/awsv2.
• Trim the CI compatibility matrix to the currently-patched OpenSearch set (2.19.x and 3.x) per the 12-month support policy; older lines (1.3.x - 2.18.x) are no longer part of the tested matrix and the 4.x client remains their supported path. No client code change (#856)
• BREAKING: Module path is now github.com/opensearch-project/opensearch-go/v5. Update import paths from /v4 to /v5; the in-source opensearchapi.X package qualifier is unchanged
• BREAKING: The code-generated API package is now the canonical opensearchapi/, replacing the hand-written v4 package (formerly previewed at v5preview/opensearchapi/). Req/Resp/Params types are fully typed and generated from the OpenAPI spec. See opensearchapi/UPGRADING_V4_TO_V5.md for the field-level delta (DocumentID -> ID, optional Params becoming *Params, shared parameters moving into embedded TimeoutParams/DebugParams, BulkResp.Items becoming []BulkItem) (#650)
• BREAKING: The default Router is now on by default. opensearchapi.NewClient/NewDefaultClient, opensearch.NewClient, and opensearchtransport.New inject opensearchtransport.NewDefaultRouter (and enable on-start discovery) unless OPENSEARCH_GO_ROUTER=false. In v4 the router was opt-in via OPENSEARCH_GO_ROUTER=true (#816)
• BREAKING: Partial-failure errors are now reported by default. Config.Errors == nil resolves to errmask.Empty (report every partial-failure category) instead of v4's errmask.All (mask everything). Set Errors: errmask.New(errmask.All) or OPENSEARCH_GO_ERROR_MASK to restore v4-style masking (#816)
• BREAKING: cmd/osgen now treats the OpenSearch plugin acronyms ISM, KNN, LTR, ML, PPL, SM, UBI, and WLM as initialisms, so generated identifiers are all-uppercase per Go convention (matching the existing API, HTTP, JSON, etc. handling). Renames every affected *_gen.go type, path builder, and method, e.g. IsmPolicy -> ISMPolicy, KnnStats -> KNNStats, SmPolicy -> SMPolicy. Update any direct references to the renamed identifiers; the lowercase plugin package names (ism, knn, ...) are unchanged (#863)
• cmd/osgen recognizes more acronyms as initialisms, extending the set from #863. Adds BM25, CJK, CSV, DFI, DFR, FS, GC, HDR, HTML, IB, ICU, IDs, JVM, LMD, LMJ, MMap, NIO, PITs, SMTP, SNS, TFIDF, UAX, WKT, and XY to the generator's acronym table, so generated identifiers use idiomatic Go capitalization (e.g. IndicesIndexSettingsSimilarityBM25, CommonAnalysisCJKAnalyzer, CommonQueryDSLIDsQuery, GetAllPITsReq, NodesStatsLastGC, ClusterStatsClusterJVM). Plural acronyms keep a lowercase s (IDs, PITs) per Go convention; the memory-map and NIO store types follow MMap/NIO. JSON wire tags are unchanged (#961)
• cmd/osgen derives clearer names for oneOf/anyOf union branches. Aggregation-result branches that the spec titles with terse codes split into readable Go names (Lterms -> LTerms, Tdigest -> TDigest, and the full long/string/unsigned/unmapped/significant family), so the As*/New*From* methods on types like SearchResultAggregationsValue read as AsLTerms/AsTDigestPercentiles. Inline-object branches carry their spec title where one exists (a hyphenated title normalizes to PascalCase, score-ranker-processor -> ScoreRankerProcessor); untitled members are named from their content instead of their spec-array position, so a branch is named for its first required field or, when it declares none, for its property names (IndicesOpenRespBodyObject0/Object1 -> ...Task/...Acknowledged, WLMQueryGroupRespResourceLimitsObject0/Object1 -> ...Memory/...CPU, SearchBodySourceObject1 -> ...ExcludesIncludes). The name no longer shifts when the spec reorders a union's members. Two branches with identical fields that cannot be told apart keep a positional ObjectN name. Branch accessors and constructors no longer repeat the union prefix, so NewInsightsSourceSourceFromInsightsSourceSourceObject1 becomes NewInsightsSourceSourceFromExcludesIncludes (#961)
• BREAKING: opensearchtransport.ConnectionObserver interface gained an OnAddressRewrite(AddressRewriteEvent) method for the new address resolver feature (embedders of BaseConnectionObserver are unaffected) (#822)
• BREAKING: Rename opensearchtransport.Client to opensearchtransport.Transport so the type name reflects its role (HTTP round-trip concerns: connection pool, retries, node selection, discovery) rather than colliding conceptually with opensearch.Client and opensearchapi.Client. The Client name is removed; update references to Transport (#853)
• BREAKING: opensearch.Request interface signature changed from GetRequest() (*http.Request, error) to GetRequest(method string) (*http.Request, error). The HTTP method is now caller-provided rather than hardcoded per operation, enabling correct method selection for operations that support multiple HTTP methods (e.g. search supports both GET and POST). This only affects code that implements or calls GetRequest directly; standard usage through client methods (e.g. client.Search(ctx, req)) is unaffected (#650)
• Bump CI and developer guide OpenSearch versions: compatibility matrix to 2.19.5, default integration test version to 3.6.0 (#810)
• Include _nodes.failures detail in discovery error messages for diagnosing intermittent CI failures on older OpenSearch versions (#823)
• Test against Opensearch 3.6.0 (#817)
• Consolidate test utilities into two canonical packages: opensearchtransport/testutil (env helpers, polling, version comparison) and opensearchapi/testutil (client-dependent helpers, test suite, JSON comparison)
• Rename singleConnectionPool to singleServerPool and statusConnectionPool to multiServerPool for clarity (#786)
• Refactor Client struct to use embedded mutex pattern for improved thread safety (#775)
• Refactor metrics struct to use atomic counters for lock-free request/failure tracking (#776)
• Test against Opensearch 2.19.4, 3.1, 3.3, and 3.4 (#782)
• Migrate all test files to context-aware API calls for proper timeout and cancellation support
• Add cluster readiness validation and improve cluster error diagnostics
• Update Docker cluster management to add version-aware role detection (cluster_manager vs master)
• Generate unique document IDs in tests for parallel test execution and eliminate known test flakes
• Reduce integration test timeout from 1h to 10m per package with parallel execution support
• Refactor transport code for improved maintainability (rename ErrInvalidRole -> InvalidRoleError, add response body cleanup, simplify initialization)
• BREAKING: Change CatTemplatesReq.Templates and IndexTemplateGetReq.IndexTemplates from []string to string to match the OpenSearch API specification, which types these path parameters as scalar name patterns (not comma-separated lists). This breakage will show up at compile time as a type mismatch and is easy to fix. Callers passing a single pattern only need to remove the slice literal (e.g. []string{"*"} becomes "*"). Callers that relied on the old behavior of joining multiple patterns can use strings.Join(patterns, ",") to produce the comma-separated string themselves.
• BREAKING: Enhanced node discovery to match OpenSearch server behavior (#765)
  • Dedicated cluster manager nodes are now excluded from client request routing by default (best practice)
  • Node selection logic now matches Java client NodeSelector.SKIP_DEDICATED_CLUSTER_MASTERS behavior
• BREAKING: Add context support to discovery and client lifecycle management
  • opensearchtransport.Discoverable interface now requires context.Context parameter: DiscoverNodes(ctx context.Context) error
  • opensearch.Client.DiscoverNodes() and opensearchtransport.Transport.DiscoverNodes() now require context.Context parameter
  • opensearch.Config and opensearchtransport.Config now accept optional Context and CancelFunc fields
  • opensearchutil.BulkIndexerConfig now accepts optional Context and CancelFunc fields
  • Enables proper context propagation for timeouts, cancellation, and graceful shutdown
  • Role compatibility validation prevents conflicting role assignments (master+cluster_manager, warm+search)
  • OpenSearch 3.0+ searchable snapshots now use warm role instead of deprecated search role
• BREAKING: Remove the signer/aws package. Use signer/awsv2, whose name mirrors AWS's own SDK-version nomenclature. For callers on released v4 this is a full AWS SDK v1 -> v2 signer migration: the constructor input changes from session.Options to aws.Config, the return type becomes the signer.Signer interface, and the removed OpenSearchService/OpenSearchServerless constants become the "es"/"aoss" literals. See UPGRADING_V5.md and USER_GUIDE.md.
• BREAKING: Replace []json.RawMessage with typed []BulkByScrollFailure for Failures field in DocumentDeleteByQueryResp, UpdateByQueryResp, and ReindexResp (#797). This is a compile-time change only -- callers that were not accessing .Failures are unaffected, and callers that were manually unmarshaling json.RawMessage can now access typed fields directly.
• Replace inline _shards struct with ResponseShards in IndexResp, DocumentCreateResp, DocumentDeleteResp, UpdateResp, IndicesRefreshResp, and IndicesCountResp to expose shard Failures and Skipped fields (#797). Code accessing resp.Shards.Total, resp.Shards.Successful, or resp.Shards.Failed compiles unchanged.
• Add omitempty to all deprecated _type JSON tags so empty values are omitted during marshaling
• Modernize tests to use Go 1.25's WaitGroup.Go() (#834)
• Make opensearch.Response.String() non-consuming for responses returned by Client.Do: Do buffers the response payload into rawBody (for both success and error responses in the default buffered mode), and String() renders from those bytes without touching Body. The receiver is a value receiver, so both Response and *Response satisfy fmt.Stringer. For an unbuffered Body (streamed responses or a hand-built Response), String() reads Body once and caches the bytes so repeat calls are consistent, but a value receiver cannot restore the caller's Body field, so that single-use stream is consumed (#859)

Deprecated

• Mark opensearchtransport.Transport.Perform and the opensearch.Client.Perform passthrough as deprecated; both remain fully functional in v4 (still buffering the response body via io.ReadAll + NopCloser) and will be removed in a future major version. New code should call opensearch.Do[T] for typed, decoded results or opensearchtransport.Transport.Stream / opensearch.Client.Stream for raw byte forwarding.
• Mark Client.Do() with a Deprecated doc annotation in favor of opensearch.Do[T]() for compile-time pointer safety; Client.Do() remains fully functional and will not be removed, but staticcheck SA1019 will nudge cross-package callers toward the safer generic alternative
• Mark opensearch.ToPointer as deprecated; it remains fully functional but will be removed in a future major version. Once the module's go directive moves to 1.26, callers can drop the helper entirely in favor of native new(value) literal syntax (e.g. new(false))

Removed

• Remove deprecated (*opensearch.Client).Perform and (*opensearchtransport.Transport).Perform; Stream(*http.Request) (*http.Response, error) is now the sole method on opensearchtransport.Interface. Custom transport implementations must implement Stream instead of Perform. The opensearch.Streamer opt-in interface and opensearch.ErrTransportMissingMethodStream sentinel are removed. (#872)
• BREAKING: Remove the EnableMetrics config flag from opensearch.Config and opensearchtransport.Config. The detailed-metrics snapshot (per-connection enumeration, per-policy breakdowns, and router cache state) is now always available; Metrics() returns the full snapshot unconditionally. The flag's only remaining purpose after #891 was to gate the detailed path, which now does its work lazily and lock-free at call time and so costs nothing until Metrics() is called. Delete any EnableMetrics field from your config (it is a compile error otherwise); see UPGRADING_V5.md. (#892)
• Remove backport.yml and dependabot_pr.yml as we are not using backport app anymore
• Stop emitting opensearchapi.Client sub-client fields that have no operations routed to them. cmd/osgen now emits a sub-client only when at least one operation targets it, dropping the previously-empty Script, ComponentTemplate, IndexTemplate, Template, and DataStream fields. Index-template and data-stream operations are reached through client.Indices.* (e.g. client.Indices.PutIndexTemplate, client.Indices.CreateDataStream); stored-script operations remain top-level on Client

Fixed

• Fix cmd/osgen silently dropping typed structs on Go type-name collisions, and add a completeness guard so future collisions fail generation instead of degrading output. Two distinct spec schemas that derived the same Go name were reduced to one by the type registry, dropping the other to raw json.RawMessage (or mis-typing a field): the search profile container collided with the per-search SearchProfile item, and multiple response bodies in the flow_framework.common and security_analytics.findings groups all derived <Group>Resp. Colliding refs are now disambiguated via documented override tables (typeNameCollisions / respTypeNameCollisions), and a panic-guard requires any new colliding ref to be enumerated. Regenerating restores typed responses for the affected operations (e.g. flow_framework search/search_state, security_analytics search_finding_correlations, and the search response's profile field), and a response schema also referenced structurally (a search hit's _source) is now emitted as a standalone type instead of dangling (#989)
• Fix plugin dispatch methods discarding the transport response, leaving Inspect().Response nil on every typed plugin response. The generated plugin dispatch template dropped the *opensearch.Response returned by request() (if _, err := request(...)); it now assigns it (resp.response, err = request(...)), matching the core client. Also serialize the discovery-path warmup recalculation and ready-list partitioning in createOrUpdateMultiNodePoolWithLock under the pool write lock -- recalculateWarmupParamsWithLock/getWarmupParamsWithLock (renamed to reflect the requirement) and the mu.activeCount write touched mu-guarded fields without holding pool.mu, racing resurrectWithLock (follow-up to #981) (#989)
• Fix node discovery hijacking the request stream with unverified, unreachable discovered nodes and masking the user-supplied seed-URL fallback. When discovered publish_address values are unroutable from the client (NAT'd or misconfigured clusters, e.g. a Kubernetes stack cluster in CI), a freshly discovered but never-health-checked node could be served to requests as a zombie -- failing every request with no route to host -- instead of returning ErrNoConnections and cascading to the reachable seed URL. Connections are now considered available for routing only when they are a user-supplied seed (assumed reachable) or a discovered node confirmed reachable, and every routing policy and pool (round-robin, role, coordinator, index/doc router, single-server, and multi-server pools) consistently honors that gate on both the enabled-bit and connection-selection paths, so the seed fallback serves requests until a discovered node health-checks clean (#952, #954, #956)
• Fix a data race on the multi-server pool's warmupRounds, warmupSkipCount, and activeListCap fields when two concurrent DiscoverNodes calls drive RolePolicy.DiscoveryUpdate on a shared transport. RolePolicy called recalculateWarmupParams (which writes those fields) without holding the pool write lock, while the roundrobin and cluster_coordinator policies took the lock for the identical call. RolePolicy.DiscoveryUpdate now computes the projected pool size and recalculates the warmup parameters under pool.Lock(), matching the other callers
• Fix a data race (reported by the race detector in TestClientCustomTransport) between multiServerPool.snapshot() and node discovery: snapshot() read activeListCap after releasing the pool read lock, while recalculateWarmupParams writes it under the write lock during DiscoveryUpdate. activeListCap, warmupRounds, and warmupSkipCount were guarded by the pool lock only by convention (declared at the top level of the struct), which let the unlocked read look correct; they are now nested inside the pool's lock-guarded mu struct so every access is spelled cp.mu.<field> and the guard is structural, and snapshot() reads activeListCap while holding the read lock. For the same reason healthCheck is moved under mu (it is rewritten by updateConnectionPool on pool reuse); this also surfaced one discovery-path read of healthCheck that had escaped the lock, now taken under the read lock
• Cache credentials in the signer/awsv2 constructors. A raw CredentialsProvider is wrapped in an aws.CredentialsCache (an already-cached provider, such as one from config.LoadDefaultConfig, is left as-is), so SigV4 signing no longer calls Credentials.Retrieve on every request. For STS-backed providers (assume-role, web identity, IRSA) the previous behavior was a per-request STS call that could exhaust the account's STS rate limits under load. signer/awsv2 shipped without this in v4.6.0.
• Fix cmd/osgen silently dropping a response struct when a response schema has a oneOf/anyOf field whose parent-scoped union name collides with the parent struct's own Go name. The union registered first and the parent struct was then dropped by the type registry (its name already taken), degrading the response to raw json.RawMessage. Such a union is now re-keyed by its referenced schema so the parent struct survives. The generator also reports any remaining Go type name collisions to stderr at generation time instead of dropping types silently. Regenerating fixes two type families: tasks.list, tasks.cancel, and delete_by_query_rethrottle change from raw Body json.RawMessage to typed structs (NodeFailures, TaskFailures, Nodes map[string]TasksTaskExecutingNode, Tasks *TasksTaskInfos), and the _common.mapping___DynamicTemplate.mapping field becomes typed *CommonMappingProperty (accounting for the large unions_gen.go/indices-put_mapping_gen.go churn). (#890)
• Fix cmd/osgen degrading two more schema shapes to raw json.RawMessage: an OpenAPI 3.1 nullable scalar (type: ["null", "<primitive>"]) fell through because kin-openapi's Type.Is matches only single-element type sets, and a response whose component schema is a bare $ref alias (Foo: {$ref: Bar}) missed the registry lookup under its alias key. Nullable scalars now resolve to the pointer primitive (*string/*int/*bool/*float64), clearing the CAT *Record cluster, and alias responses follow the $ref chain to the registered struct, fixing ISM add/delete/get/remove_policy + retry_index and the seven ml.search_* responses. (#890)
• Fix cmd/osgen generating a phantom request body for the _sql/stats and _ppl/stats POST operations. The server (RestSqlStatsAction/RestPPLStatsAction) ignores the request body, so the spec's body schema is a defect; removing it drops the dead SQLStats type. The typed client no longer sends a body to these endpoints. (#890)
• Fix BulkIndexer OnFailure nil pointer dereference when reading BulkRespItem.Error on status-only failures (e.g. HTTP 404 without an error object) or transport-level flush errors by ensuring callbacks always receive a non-nil Error (#679)
• Generate query parameters whose value 0 is meaningful as *int instead of int so a deliberate 0 reaches the wire. These params previously used the != 0 emission guard shared by all integer params, which silently dropped a deliberate 0 -- breaking optimistic-concurrency writes with if_seq_no=0 (the sequence number of the first document written to a shard) and search size=0 (aggregations with no hits). cmd/osgen now promotes such params to *int with a nil guard, mirroring the existing *bool treatment. The promotion is scoped per operation (currently if_seq_no/if_primary_term on delete/index/update and the plugin policy writes ism.put_policy/ism.put_policies/rollups.put/sm.update_policy/transforms.put, plus size on search), since the same wire name is a page-size with no meaningful 0 on other operations. The core _create operation does not accept if_seq_no/if_primary_term, so it is intentionally excluded
• Fix BulkIndexerStats.NumAdded overcounting items rejected by Add() when the caller's context is cancelled before the item could be enqueued: increment NumAdded only after the queue accepts the item, and add a new BulkAddFailCount counter for items dropped on the <-ctx.Done() branch. Migrate bulkIndexerStats fields to sync/atomic.Uint64 typed values so future direct access is a compile-time error rather than a -race-only finding (#783)
• Fix opensearchtransport.Transport.setReqGlobalHeader comparing the per-request header value against the global header name, so a request-level header never suppressed the matching global default and both were sent (#859)
• Fix gzip buffer-pool nil poisoning on compress error: gzipCompressor.compress returned (nil, err) while the caller's deferred collectBuffer still ran, putting a typed-nil *bytes.Buffer into the sync.Pool that panics on the next Get().Reset() (#859)
• Fix opensearchtransport.Transport.Perform silently dropping io.ReadAll errors during response buffering via := shadowing; the read error now propagates wrapped in the new opensearchtransport.ErrResponseBodyRead sentinel, and opensearch.Client.Do classifies the (resp != nil, err != nil) case via errors.Is so only genuine body-read failures are labeled ErrReadBody (an unrelated transport error returned alongside a response, such as context cancellation during retry backoff, is no longer misreported as a read failure). As a consequence, opensearch.Client.Do now returns a non-nil *Response alongside a non-nil error in this case where it previously returned (nil, err); callers detecting a hard transport failure should check resp == nil rather than err != nil (#859)
• Fix error-response body not being closed in opensearch.ParseError. ParseError now closes the original body before re-wrapping the read bytes in a NopCloser. The generated opensearchapi do() no-decode error path no longer needs its own drain: opensearch.Do routes through the buffered opensearchtransport.Transport.Perform, so the returned resp.Body is already an in-memory NopCloser over the full payload and stays readable for the caller (#859)
• Fix response-body lifecycle on the raw RoundTrip paths that lack Perform's buffering safety net, where closing a partially-read body defeated HTTP keep-alive: the stats poller (cluster_health.go), discovery's /_cat/shards, /_cluster/state/metadata, and /_nodes paths, and the fetchClusterHealth/baselineHealthCheck/hardwareInfoHealthCheck pollers now drain to EOF (io.Copy(io.Discard, ...)) before close (covering both non-200 returns and json.Decode success paths that stop before EOF). The AWS v1 and v2 signers now close the request body on the read-error path in hexEncodedSha256OfRequest (#859)
• Fix Client.Do to buffer every response body into rawBody -- decoded success, error, and no-decode (nil dataPointer) success alike (previously some paths, including the nil-dataPointer success path, were left unbuffered). Without this, a value-receiver Response.String() (e.g. log.Printf("%s", resp)) drained the single-use Body and left a subsequent ParseError reading an empty payload (surfacing ErrJSONUnmarshalBody instead of the real API error). Now String() renders from rawBody without touching Body and ParseError reads an intact body (#859)
• Add typed response-format defaults for generated opensearchapi/ cat, list, ppl, and sql operations: when the caller leaves Format unset, the SDK now emits the value the typed Resp struct expects (json for cat/list/explain, jdbc for ppl/sql query) instead of letting the server fall back to a default the JSON decoder cannot handle.
• Replace WaitForAllNodesReady inline require.Eventually loop with a layered readiness FSM (internal/test/readiness) that observes per-node progression through LayerTCP -> LayerHTTP -> LayerClusterJoin -> LayerStatsReady, records transitions including regressions, and emits a structured per-node diagnostic with the full last cat-nodes response on timeout. Per-layer budgets are tuned for CI pessimism (cold JVM startup is the long pole); total budget for TargetClusterReady is 6.5 minutes. (#650)
• Fix bulk indexer HTML-escaping _id and routing values containing <, >, or & characters, causing OpenSearch to store escaped values (e.g., \u003croot_account\u003e stored instead of <root_account>), leading to duplicate documents, unreachable data on read-by-ID paths, and potential shard routing mismatches. Present since the json.Marshal migration in 2021 (commit 3da59092). Replace json.Marshal with json.NewEncoder + SetEscapeHTML(false) in opensearchutil.worker.writeMeta and opensearchutil.JSONReader; replace per-worker aux []byte with sync.Pool-backed *bytes.Buffer; add table-driven test coverage for writeMeta edge cases and refactor remaining TestBulkIndexer subtests to table-driven require-based style (#824)
• Fix pool replacement orphaning resurrection goroutines during node discovery, causing connections to become permanently dead with no active health checker (#786)
• Fix multi-to-single pool demotion leaking resurrection goroutines by giving each multiServerPool its own derived context and cancelling it on demotion (#830)
• Extract newMultiServerPoolFromClientWithLock as single source of truth for Client-to-pool settings propagation (#786)
• Skip shard routing integration tests on OpenSearch < 2.2.0 with security plugin due to server-side OptionalDataException from non-thread-safe User serialization (opensearch-project/security#1970)
• Fix URL path construction across 74 GetRequest methods where empty path segments produced a double-slash // that http.NewRequest misparsed as an RFC 3986 authority separator; replace manual strings.Builder paths with typed path builder structs that reject empty required segments (#617, #650)
• Eliminate per-request url.Parse overhead by constructing *http.Request directly with a coalesced struct; reduce per-request allocations from 8/2930B to 2/472B for typical operations (#650)
• Fix alias, mapping, settings, and block API URL path construction when Indices is empty, which caused http.NewRequest to misparse the double-slash as an authority separator (#650)
• Fix discovery pool wipe when all cluster nodes time out during /_nodes/http fan-out: parse _nodes metadata envelope and return errDiscoveryEmpty when successful == 0, preserving the existing connection pool for retry (#821)
• Skip shard routing integration tests on OpenSearch < 2.2.0 with security plugin due to server-side OptionalDataException from non-thread-safe User serialization (opensearch-project/security#1970)
• Fix flaky TestDefaultHealthCheck_RetryAfterMaxRetry: replace wall-clock time.Sleep + atomic.Int64 synchronization with context cancellation (ctx.Done()), and widen maxRetryClusterHealth to 5s so the baseline HTTP round-trip cannot race past the retry interval (#787)
• Skip opensearchtransport integration tests on OpenSearch < 2.2.0 with security plugin due to server-side OptionalDataException from non-thread-safe User serialization (opensearch-project/security#1970)
• Skip shard routing integration tests on OpenSearch < 2.2.0 with security plugin due to server-side OptionalDataException from non-thread-safe User serialization (opensearch-project/security#1970)
• Fix connection lifecycle bug in multiServerPool.OnFailure where connections were scheduled for resurrection before being moved from ready to dead list, causing potential race conditions
• Fix flaky connection integration test by replacing arbitrary sleep times with proper server readiness polling
• Fix cluster readiness checks in integration tests to handle HTTPS cold start delays (increase timeout to 15s)
• Fix GitHub Actions workflow authentication for OpenSearch 2.12.0+ password changes (admin -> myStrongPassword123!)
• Fix Docker cluster management to properly handle version-specific configurations and clean stale images/volumes
• Fix OpenSearch 2.8.0+ Tasks API compatibility by adding cancellation_time_millis field to TasksListTask struct
• Fix OpenSearch 3.1.0+ API compatibility by adding phase_results_processors field to nodes API and time_in_execution fields to cluster pending tasks API
• Fix OpenSearch 3.2.0+ API compatibility by adding max_last_index_request_timestamp and startree query fields across nodes stats, indices stats, and cat APIs, plus settings field to security plugin health API
• Fix OpenSearch 3.3.0+ API compatibility by adding neural_search breaker, query_failed and startree_query_failed search fields, search pipeline system_generated fields across multiple APIs, plus ingestion_status field to cluster state API and jwks_uri field to security config API
• Fix OpenSearch 3.4.0+ API compatibility by adding warmer fields to merges section, parallelism field to thread pool, and status_counter field across multiple APIs
• Fix cat indices API field naming compatibility across OpenSearch versions by using forward-compatible field names (PrimarySearchStartreeQuery) that match the corrected 3.3.0+ naming, with fallback support for the temporary 3.2.0 field names
• Fix cat APIs data type compatibility by changing byte fields from int to string to properly handle values like "0b"
• Fix floating point precision loss in nodes stats concurrent_avg_slice_count field by changing from float32 to float64
• Fix ISM RefreshSearchAnalyzers missing leading slash in URL path, causing HTTP/2 request failures (#686)
• Default the benchmark pprof server to an ephemeral loopback port so back-to-back go test -bench runs no longer collide on a TIME_WAIT socket held by the prior run. The startup logic moves into an internal/pprofutil package that registers the pprof handlers on a private mux (off http.DefaultServeMux); PPROF_ADDR pins an explicit host:port when needed. (#864)

Security

Dependencies

• Bump golangci-lint from v2.11.2 to v2.11.4
• Bump golang.org/x/sync from v0.19.0 to v0.20.0 (#831)
• Bump golang.org/x/mod from v0.33.0 to v0.35.0 (#831)
• Bump github.com/wI2L/jsondiff from v0.7.0 to v0.7.1 (#831)
• Bump github.com/aws/aws-sdk-go-v2 from v1.41.1 to v1.41.7 (#831)
• Bump github.com/aws/aws-sdk-go-v2/config from v1.32.7 to v1.32.17 (#831)
• Bump github.com/aws/aws-sdk-go-v2/credentials from v1.19.7 to v1.19.16 (#831)
• Bump github.com/aws/smithy-go from v1.24.0 to v1.25.1 (#831)
• Bump github.com/aws/aws-sdk-go-v2/config from 1.32.6 to 1.32.7 (#767)

4.6.0

Dependencies

• Bump github.com/aws/aws-sdk-go-v2/config from 1.29.14 to 1.32.5 (#707, #711, #719, #730, #737, #761)
• Bump github.com/aws/aws-sdk-go-v2 from 1.36.4 to 1.41.0 (#710, #720, #759)
• Bump github.com/stretchr/testify from 1.10.0 to 1.11.1 (#728)
• Bump github.com/aws/aws-sdk-go from 1.55.7 to 1.55.8 (#716)
• Bump go version from 1.24.0 to 1.25.9 in order to resolve certain CVEs. Details in the Pull Request (#825)

Added

• Adds new fields for Opensearch 3.0 (#702)
• Allow users to override signing port (#721)
• Add phase_took features supported from OpenSearch 2.12 (#722)
• Adds the action to refresh the search analyzers to the ISM plugin (#686)

Changed

• Test against Opensearch 3.0 (#702)
• Add more SuggestOptions to SearchResp (#713)
• Updates Go version to 1.24 (#674)
• Replace golang.org/x/exp/slices usage with built-in slices (#674)
• Update golangci-linter to 1.64.8 (#740)
• Change MaxScore to pointer (#740)
• Update workflow action (#760)
• Migrate to golangci-lint v2 (#760)

Deprecated

Removed

Fixed

• Missing "caused by" information in StructError (#752)
• Add missing ignore_unavailable, allow_no_indices, and expand_wildcards params to MSearch (#757)
• Fix UpdateResp to correctly parse the get field when _source is requested in update operations. (#739)

Security

4.5.0

Dependencies

• Bump github.com/aws/aws-sdk-go-v2/config from 1.29.6 to 1.29.14 (#692)
• Bump github.com/aws/aws-sdk-go from 1.55.6 to 1.55.7 (#696)
• Bump github.com/wI2L/jsondiff from 0.6.1 to 0.7.0 (#700)

Added

• Adds DataStream field to IndicesGetResp struct (#701)
• Adds InnerHits field to SearchResp (#672)
• Adds FilterPath param (#673)
• Adds Aggregations field to MSearchResp (#690)

Changed

• Bump golang version to 1.22 (#691)
• Change ChangeCatRecoveryItemResp Byte fields from int to string (#691)
• Changed log formatted examples code (#694)
• Improve the error reporting of invalid body response (#699)

Deprecated

Removed

Fixed

Security

4.4.0

Added

• Adds Highlight field to SearchHit (#654)
• Adds MatchedQueries field to SearchHit (#663)
• Adds support for Opensearch 2.19 (#668)

Changed

Deprecated

Removed

Fixed

Security

Dependencies

• Bump github.com/aws/aws-sdk-go from 1.55.5 to 1.55.6 (#657)
• Bump github.com/wI2L/jsondiff from 0.6.0 to 0.6.1 (#643)
• Bump github.com/aws/aws-sdk-go-v2 from 1.32.2 to 1.36.1 (#664)
• Bump github.com/stretchr/testify from 1.9.0 to 1.10.0 (#644)
• Bump github.com/aws/aws-sdk-go-v2/config from 1.27.43 to 1.29.6 (#665)

4.3.0

Added

• Adds ISM Alias action (#615)
• Adds support for opensearch 2.17 (#623)

Changed

Deprecated

Removed

Fixed

• Fix ISM Transition to omitempty Conditions field (#609)
• Fix ISM Allocation field types (#609)
• Fix ISM Error Notification types (#612)
• Fix signer receiving drained body on retries (#620)
• Fix Bulk Index Items not executing failure callbacks on bulk request failure (#626)

Security

Dependencies

• Bump github.com/aws/aws-sdk-go-v2/config from 1.27.31 to 1.27.43 (#611, #630, #632)
• Bump github.com/aws/aws-sdk-go-v2 from 1.32.1 to 1.32.2 (#631)

4.2.0

Dependencies

• Bump github.com/aws/aws-sdk-go-v2/config from 1.27.23 to 1.27.31 (#584, #588, #593, #605)
• Bump github.com/aws/aws-sdk-go from 1.54.12 to 1.55.5 (#583, #590, #595, #596)

Added

• Adds Suggest to SearchResp (#602)
• Adds MaxScore to ScrollGetResp (#607)

Changed

• Split SnapshotGetResp into sub structs (#603)

Deprecated

Removed

• Remove workflow tests against gotip (#604)

Fixed

Security

4.1.0

Added

• Adds the Routing field in SearchHit interface. (#516)
• Adds the SearchPipelines field to SearchParams (#532)
• Adds support for OpenSearch 2.14 (#552)
• Adds the Caches field to Node stats (#572)
• Adds the SeqNo and PrimaryTerm fields in SearchHit (#574)
• Adds guide on configuring the client with retry and backoff (#540)
• Adds OpenSearch 2.15 to compatibility workflow test (#575)

Changed

• Security roles get response struct has its own sub structs without omitempty (#572)

Deprecated

Removed

Fixed

• Fixes empty request body on retry with compression enabled (#543)
• Fixes Conditions in PolicyStateTransition of ISM plugin (#556)
• Fixes integration test response validation when response is null (#572)
• Adjust security Role struct for FLS from string to []string (#572)
• Fixes wrong response parsing for indices mapping and recovery (#572)
• Fixes wrong response parsing for security get requests (#572)
• Fixes opensearchtransport ignores request context cancellation when retryBackoff is configured (#540)
• Fixes opensearchtransport sleeps unexpectedly after the last retry (#540)
• Improves ParseError response when server response is an unknown json (#592)

Security

Dependencies

• Bump github.com/aws/aws-sdk-go from 1.51.21 to 1.54.12 (#534, #537, #538, #545, #554, #557, #563, #564, #570, #579)
• Bump github.com/wI2L/jsondiff from 0.5.1 to 0.6.0 (#535, #566)
• Bump github.com/aws/aws-sdk-go-v2/config from 1.27.11 to 1.27.23 (#546, #553, #558, #562, #567, #571, #577)
• Bump github.com/aws/aws-sdk-go-v2 from 1.27.0 to 1.30.1 (#559, #578)

4.0.0

Added

• Adds GlobalIOUsage struct for nodes stats (#506)
• Adds the Explanation field containing the document explain details to the SearchHit struct. (#504)
• Adds new error types (#512)
• Adds handling of non json errors to ParseError (#512)
• Adds the Failures field to opensearchapi structs (#510)
• Adds the Fields field containing the document fields to the SearchHit struct. (#508)
• Adds security plugin (#507)
• Adds security settings to container for security testing (#507)
• Adds cluster.get-certs to copy admin certs out of the container (#507)
• Adds the Fields field containing stored fields to the DocumentGetResp struct (#526)
• Adds ism plugin (#524)

Changed

• Uses docker compose v2 instead of v1 (#506)
• Updates go version to 1.21 (#509)
• Moves Error structs from opensearchapi to opensearch package (#512)
• Moves parseError function from opensearchapi to opensearch package as ParseError (#512)
• Changes ParseError function to do type assertion to determine error type (#512)
• Removes unused structs and functions from opensearch (#517)
• Adjusts and extent opensearch tests for better coverage (#517)
• Bumps codecov action version to v4 (#517)
• Changes bulk error/reason field and some cat response fields to pointer as they can be nil (#510)
• Adjust workflows to work with security plugin (#507)
• Updates USER_GUIDE.md and add samples (#518)
• Updates opensearchtransport.Client to use pooled gzip writer and buffer (#521)
• Use go:build tags for testing (#52?)

Deprecated

Removed

Fixed

• Fixes search request missing a slash when no indices are given (#470)
• Fixes opensearchtransport check for nil response body (#517)

Security

Dependencies

• Bumps github.com/aws/aws-sdk-go-v2 from 1.25.3 to 1.26.1
• Bumps github.com/wI2L/jsondiff from 0.4.0 to 0.5.1
• Bumps github.com/aws/aws-sdk-go from 1.50.36 to 1.51.21
• Bumps github.com/aws/aws-sdk-go-v2/config from 1.27.7 to 1.27.11

3.1.0

Added

• Adds new struct fields introduced in OpenSearch 2.12 (#482)
• Adds initial admin password environment variable and CI changes to support 2.12.0 release (#449)
• Adds merge_id field for indices segment request (#488)

Changed

• Updates workflow action versions (#488)
• Changes integration tests to work with secure and unsecure OpenSearch (#488)
• Moves functions from opensearch/internal/test to opensearchutil/testutil for shared test utilities (#488)
• Changes custom_foldername field to pointer as it can be null (#488)
• Changs cat indices Primary and Replica field to pointer as it can be null (#488)
• Replaces ioutil with io in examples and integration tests #495

Fixed

• Fix incorrect SigV4 x-amz-content-sha256 with AWS SDK v1 requests without a body (#496)

Dependencies

• Bumps github.com/aws/aws-sdk-go from 1.48.13 to 1.50.36
• Bumps github.com/aws/aws-sdk-go-v2/config from 1.25.11 to 1.27.7
• Bumps github.com/stretchr/testify from 1.8.4 to 1.9.0

3.0.0

Added

• Adds Err() function to Response for detailed errors (#246)
• Adds golangci-lint as code analysis tool (#313)
• Adds govulncheck to check for go vulnerablities (#405)
• Adds opensearchapi with new client and function structure (#421)
• Adds integration tests for all opensearchapi functions (#421)
• Adds guide on making raw JSON REST requests (#399)
• Adds IPV6 support in the DiscoverNodes method (#458)

Changed

• Removes the need for double error checking (#246)
• Updates and adjusted golangci-lint, solve linting complains for signer (#352)
• Solves linting complains for opensearchtransport (#353)
• Updates Developer guide to include docker build instructions (#385)
• Tests against version 2.9.0, 2.10.0, run tests in all branches, changes integration tests to wait for OpenSearch to start (#392)
• Makefile: uses docker golangci-lint, run integration test on . folder, change coverage generation (#392)
• golangci-lint: updates rules and fail when issues are found (#421)
• go: updates to golang version 1.20 (#421)
• guids: updates to work for the new opensearchapi (#421)
• Adjusts tests to new opensearchapi functions and structs (#421)
• Changes codecov to comment code coverage to each PR (#410)
• Changes module version from v2 to v3 (#444)
• Handle unexpected non-json errors with the response body (#523)

Deprecated

• Deprecates legacy API /_template (#390)

Removed

• Removes all old opensearchapi functions (#421)
• Removes /internal/build code and folders (#421)

Fixed

• Corrects AWSv4 signature on DataStream Stats with no index name specified (#338)
• Fixes GetSourceRequest Source field and deprecated the Source parameter (#402)
• Corrects developer guide summary with golang version 1.20 (#434)

Dependencies

• Bumps github.com/aws/aws-sdk-go from 1.44.263 to 1.48.13
• Bumps github.com/aws/aws-sdk-go-v2 from 1.18.0 to 1.23.5
• Bumps github.com/aws/aws-sdk-go-v2/config from 1.18.25 to 1.25.11
• Bumps github.com/stretchr/testify from 1.8.2 to 1.8.4
• Bumps golang.org/x/net from 0.7.0 to 0.17.0
• Bumps github.com/golangci/golangci-lint-action from 1.53.3 to 1.54.2

2.3.0

Added

• Adds implementation of Data Streams API (#257)
• Adds Point In Time API (#253)
• Adds InfoResp type (#253)
• Adds markdown linter (#261)
• Adds testcases to check upsert functionality (#269)
• Adds @Jakob3xD to co-maintainers (#270)
• Adds dynamic type to _source field (#285)
• Adds testcases for Document API (#285)
• Adds index_lifecycle guide (#287)
• Adds bulk guide (#292)
• Adds search guide (#291)
• Adds document_lifecycle guide (#290)
• Adds index_template guide (#289)
• Adds advanced_index_actions guide (#288)
• Adds testcases to check UpdateByQuery functionality (#304)
• Adds additional timeout after cluster start (#303)
• Adds docker healthcheck to auto restart the container (#315)

Changed

• Uses []string instead of string in SnapshotDeleteRequest (#237)
• Updates workflows to reduce CI time, consolidate OpenSearch versions, update compatibility matrix (#242)
• Moves @svencowart to emeritus maintainers (#270)
• Reads, closes and replaces the http Response Body (#300)

Fixed

• Corrects curl logging to emit the correct URL destination (#101)

Dependencies

• Bumps github.com/aws/aws-sdk-go from 1.44.180 to 1.44.263
• Bumps github.com/aws/aws-sdk-go-v2 from 1.17.4 to 1.18.0
• Bumps github.com/aws/aws-sdk-go-v2/config from 1.18.8 to 1.18.25
• Bumps github.com/stretchr/testify from 1.8.1 to 1.8.2

2.2.0

Added

• Adds Github workflow for changelog verification (#172)
• Adds Go Documentation link for the client (#182)
• Adds support for Amazon OpenSearch Serverless (#216)

Removed

• Removes info call before performing every request (#219)

Fixed

• Renames the sequence number struct tag to if_seq_no to fix optimistic concurrency control (#166)
• Fixes RetryOnConflict on bulk indexer (#215)

Dependencies

• Bumps github.com/aws/aws-sdk-go-v2 from 1.17.1 to 1.17.3
• Bumps github.com/aws/aws-sdk-go-v2/config from 1.17.10 to 1.18.8
• Bumps github.com/aws/aws-sdk-go from 1.44.176 to 1.44.180
• Bumps github.com/aws/aws-sdk-go from 1.44.132 to 1.44.180
• Bumps github.com/stretchr/testify from 1.8.0 to 1.8.1
• Bumps github.com/aws/aws-sdk-go from 1.44.45 to 1.44.132