v0.121.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.121.0 and the opentelemetry-collector-contrib v0.121.0 releases where appropriate.

🛑 Breaking changes 🛑

• (Splunk) Support for Oracle Linux 7 has been dropped (#5974)
• (Core) confighttp: Make the client config options max_idle_conns, max_idle_conns_per_host, max_conns_per_host, and idle_conn_timeout integers (#9478)
  All four options can be set to 0 where they were previously set to null
• (Contrib) awss3exporter: Replaced the s3_partition option with s3_partition_format to provide more flexibility to users. (#37915, #37503)
  Users can provide custom file partitions using strftime formatting.
  The default value of year=%Y/month=%m/day=%d/hour=%H/minute=%M matches the older pattern (with s3_partition: minute)
  If users do not provide a value for s3_prefix, the exporter will not create a / folder in the bucket.
• (Contrib) processor/k8sattributes: Move k8sattr.fieldExtractConfigRegex.disallow feature gate to stable (#25128)
• (Contrib) signalfxexporter: Remove the deprecated configuration option translation_rules (#35332)
  Please use processors to handle desired metric transformations instead. Find migration guidance in the
  translation rules migration guide.

🚩 Deprecations 🚩

• (Core) exporterhelper: Deprecate min_size_items and max_size_items in favor of min_size and max_size. (#12486)
• (Contrib) prometheusreceiver: Deprecate metric start time adjustment in the prometheus receiver. It is being replaced by the metricstarttime processor. (#37186)
  Start time adjustment is still enabled by default. To disable it, enable the | receiver.prometheusreceiver.RemoveStartTimeAdjustment feature gate.

💡 Enhancements 💡

• (Core) pipeline: output pipeline name with signal as signal[/name] format in logs. (#12410)
• (Core) memorylimiter: Add support to configure min GC intervals for soft and hard limits. (#12450)
• (Core) otlpexporter: Update the stability level for logs, it has been as stable as traces and metrics for some time. (#12423)
• (Core) service: Create a new subcommand to dump the initial configuration after resolving/merging. (#11479)
  To use the print-initial-config subcommand, invoke the Collector with the subcommand and corresponding feature gate: otelcol print-initial-config --feature-gates=otelcol.printInitialConfig --config=config.yaml.
  Note that the feature gate enabling this flag is currently in alpha stability, and the subcommand may
  be changed in the future.
• (Core) memorylimiterprocessor: Add support for profiles. (#12453)
• (Core) otelcol: Converters are now available in the components command. (#11900, #12385)
• (Core) confmap: Surface YAML parsing errors when they happen at the top-level. (#12180)
  This adds context to some instances of the error "retrieved value (type=string) cannot be used as a Conf", which typically happens because of invalid YAML documents
• (Core) pprofile: Add LinkIndex attribute to the generated Sample type (#12485)
• (Core) exporterhelper: Stabilize exporter.UsePullingBasedExporterQueueBatcher and remove old batch sender (#12425)
• (Contrib) processor/resourcedetection: Introduce retry logic for failed resource detection. (#34761)
• (Contrib) pkg/ottl: Support dynamic indexing of math expressions in maps and slices (#37644)
• (Contrib) receiver/sqlquery: Add support for SapASE (sybase) database connections (#36328)
• (Contrib) iisreceiver: Added state and uptime metrics for application pools (#34924)
• (Contrib) pkg/stanza: Add entry's timestamp and attributes to errors logs from log transformers processors (#37285)
  When a log transformer processor fails to process an log entry it will include entry's timestamp and attributes in its own logs.
  With this information the user can more easily identify the log file and find the entry that's having issues.
• (Contrib) kafkareceiver: Add error backoff configuration to kafka receiver which allows to wait and retry a failed message when the next consumer returns some errors. (#37009)
• (Contrib) receiver/kafkametricsreceiver: Add refresh_frequency config to kafkametricsreceiver, to configure custom duration for cluster metadata refresh (#37896)
  • Helps fine tuning the refresh_frequency, and enables custom cluster metadata refresh intervals
  • Default refresh_frequency is set 10 minutes from Sarama library defaults
• (Contrib) processor/resourcedetection: Add k8s.cluster.uid to kubeadm detector (#38207)
• (Contrib) mongodbreceiver: Added mongodb replica metrics and routing logic for multiple mongodb instances (#37517)
• (Contrib) pkg/ottl: Add event_index to the available paths of the span event context (#35778)
• (Contrib) pkg/ottl: Introduce Weekday() converter function (#38126)
• (Contrib) prometheusreceiver: Make use of creation timestamp from prometheus (#36473)
• (Contrib) processor/redaction: Introduce 'blocked_key_patterns' parameter (#35830)
• (Contrib) awss3exporter: Added acl option (#37935)
• (Contrib) receiver/sqlserverreceiver: Add server.address and server.port resource attributes to SQL server receiver. (#35183)
  The new resource attributes are added to the SQL server receiver to distinguish metrics coming from different SQL server instances.
  • (Contrib) server.address: The address of the SQL server host, disabled by default.
  • (Contrib) server.port: The port of the SQL server host, disabled by default.

🧰 Bug fixes 🧰

• (Core) service: Fix crash at startup when converting from v0.2.0 to v0.3.0 (#12438)
• (Core) service: fix bug in parsing service::telemetry configuration (#12437)
• (Core) exporterhelper: Fix bug where the error logged when conversion of data fails is always nil (#12510)
• (Contrib) ecsobserver: Fixed ecs task tags not being included in metadata labels (#38278)
• (Contrib) redactionprocessor: Fix redaction processor to redact span event attributes (#36633)
• (Contrib) azuremonitorreceiver: Fix bug where the time grain wasn't honored (#37337)
• (Contrib) exporter/splunk_hec: Do not pass errors from draining the response body to the pipeline as a export failure. (#38118)
• (Contrib) kafkametricsreceiver: Fix incorrect cluster admin initialization in consumer scraper (#36818)
• (Contrib) pkg/ottl: Change the `ottlmetr...

v0.120.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.120.0 and the opentelemetry-collector-contrib v0.120.1 releases where appropriate.

🛑 Breaking changes 🛑

• (Contrib) receiver/prometheus: Prometheus receiver now uses scrapers in Prometheus 3.0. (#36873)
  There are a number of breaking changes in Prometheus 3.0. Learn more about those changes and migration guide on https://prometheus.io/docs/prometheus/latest/migration/.
  As a result of adding support for UTF-8 names,
  the metrics and labels containing UTF-8 characters are no longer escaped. Consequently, the dots (.) in internal
  collector metrics and resource attributes scraped by Prometheus are no longer replaced with underscores (_).

  • The service_name, service_instance_id, and service_version resource attributes are now scraped as
    service.name, service.instance.id, and service.version, respectively.
  • The following metrics containing dots reported by several components are no longer escaped:
    • filter processor:
      • processor_filter_datapoints_filtered -> processor_filter_datapoints.filtered
      • processor_filter_logs_filtered -> processor_filter_logs.filtered
      • processor_filter_spans_filtered -> processor_filter_spans.filtered
    • deltatocumulative processor:
      • deltatocumulative_streams_tracked -> deltatocumulative.streams.tracked
      • deltatocumulative_streams_tracked_linear -> deltatocumulative.streams.tracked.linear
      • deltatocumulative_streams_limit -> deltatocumulative.streams.limit
      • deltatocumulative_streams_evicted -> deltatocumulative.streams.evicted
      • deltatocumulative_streams_max_stale -> deltatocumulative.streams.max_stale
      • deltatocumulative_datapoints_processed -> deltatocumulative.datapoints.processed
      • deltatocumulative_datapoints_dropped -> deltatocumulative.datapoints.dropped
      • deltatocumulative_datapoints_linear -> deltatocumulative.datapoints.linear
      • deltatocumulative_gaps_length -> deltatocumulative.gaps.length
    • googlecloudpubsub receiver:
      • receiver_googlecloudpubsub_stream_restarts -> receiver.googlecloudpubsub.stream_restarts

• (Contrib) activedirectorydsreceiver: Fixed typo in the attribute distingushed_names, renaming it to distinguished_names. (#37606)

• (Contrib) receiver/hostmetrics: Remove receiver.hostmetrics.normalizeProcessCPUUtilization feature gate (#34763)

• (Contrib) tailsamplingprocessor: Fix the decision timer metric to capture longer latencies beyond 50ms. (#37722)
  This changes the unit of the decision timer metric from microseconds to milliseconds.

• (Contrib) routingconnector: Remove match_once configuration parameter. (#36824)

• (Core) service: Align component logger attributes with those defined in RFC (#12217)
  See Pipeline Component Telemetry RFC

💡 Enhancements 💡

• (Splunk) Make Windows TA agnostic to Powershell ExecutionPolicy (#5935)
• (Contrib) processor/transformprocessor: Add support for global conditions and error mode overrides. (#29017)
  Global conditions are now available for context-inferred structured configurations, allowing the use of fully
  qualified paths. Additionally, a new configuration key called error_mode has been added to the context statements group.
  This key determines how the processor reacts to errors that occur while processing that specific group of statements.
  When provided, it overrides the top-level error mode, offering more granular control over error handling.
• (Contrib) pkg/stanza: Allow users to configure initial buffer size (#37786)
• (Contrib) vcenterreceiver: Adds three more vCenter virtual machine performance metrics (#37488)
• (Contrib) k8sclusterreceiver: Adds new descriptive attributes/metadata to the k8s.namespace and the container entity emitted from k8sclusterreceiver. (#37580)
  • Adds the following attributes to k8s.namespace entity:
    • k8s.namespace.phase: The phase of a namespace indicates where the namespace is in its lifecycle. E.g. 'active', 'terminating'
    • k8s.namespace.creation_timestamp: The time when the namespace object was created.
  • Adds the following attributes to container entity:
    • container.creation_timestamp: The time when the container was started. Only available if container is either in 'running' or 'terminated' state.
• (Contrib) splunkenterprisereceiver: Added a new splunk.health metric. (#36695)
• (Contrib) resourcedetectionprocessor: add the Dynatrace detector to the resource detection processor (#37577)
• (Contrib) extension/oauth2clientauth: Add expiry_buffer config to oauth2client extension, allowing token refresh before expiration with a default buffer of 5 minutes. (#35148)
  • Prevents authentication failures by refreshing the token early.
  • The default expiry buffer is set to 5 minutes, and users can adjust it as needed.
• (Contrib) googlecloudpubsubreceiver: Turn noisy warn log about Pub/Sub servers into debug, and turn the reset count into a metric (#37571)
  The receiver uses the Google Cloud Pub/Sub StreamingPull API and keeps a open connection. The Pub/Sub servers
  recurrently close the connection after a time period to avoid a long-running sticky connection. Before the
  receiver logged warn log lines everytime this happened. These log lines are moved to debug so that fleets with
  lots of collectors with the receiver don't span logs at warn level. To keep track of the resets, whenever a
  connection reset happens a otelcol_receiver_googlecloudpubsub_stream_restarts metric is increased by one.
• (Contrib) processor/redaction: Introduce 'allowed_values' parameter for allowed values of attributes (#35840)
• (Contrib) routingconnector: Avoid unnecessary copy of the data in routing connector (#37946)
• (Contrib) awscontainerinsightreceiver: Add support for HOST_PROC environment variable in AWS Container Insight Receiver. (#35862)
• (Contrib) syslogreceiver: Support setting on_error config for syslog receiver. (#36906)
• (Contrib) processor/tailsampling: Adds support for optionally recording the policy (and any composite policy) associated with an inclusive tail processor sampling decision.
  This functionality is disabled by default, you can enable it by passing the following feature flag to the collector: +processor.tailsamplingprocessor.recordpolicy
  (#35180)
• (Contrib) tailsamplingprocessor: makes the numeric_attribute more flexible and allows to set only min_value or max_value, without the need to set both (#37328)
  This is useful to have simple configurations like these:

  {
    type: numeric_attribute,
    numeric_attribute: {
      key: http.status_code,
      min_value: 400
    }
  }
  

• (Core) otlpreceiver: Update stability for logs (#12335)
• (Core) exporterhelper: Implement sync disabled queue used when batching is enabled. (#12245)
• (Core) exporterhelper: Enable the new pull-based batcher in exporterhelper (#12291)
• (Core) exporterhelper: Update queue size after the element is done exported (#12399)
  After this change the active queue size will include elements in the process of being exported.
• (Core) otelcol: Add featuregate command to display information about available features (#11998)
  The featuregate command allows users to view detailed information about feature gates
  including their status, stage, and description.

🧰 Bug fixes 🧰

• (Contrib) `...

v0.119.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.119.0 and the opentelemetry-collector-contrib v0.119.0 releases where appropriate.

🚩 Deprecations 🚩

• (Contrib) signalfxreceiver: access_token_passthrough is deprecated (#37575)
  Please use include_metadata in the receiver instead and add the following config to the batch processor:

  batch:
    metadata_keys: [X-Sf-Token]

🛑 Breaking changes 🛑

• (Splunk) Deprecated ASP.NET and .NET SignalFx monitors are removed (#5868)
• (Core) exporters: Rename exporter span signal specific attributes (e.g. "sent_spans" / "send_failed_span") to "items.sent" / "items.failed". (#12165)
• (Core) exporters: Change exporter ID to be a Span level attribute instead on each event. (#12164)

💡 Enhancements 💡

• (Core) configtls: Allow users to mention their preferred curve types for ECDHE handshake (#12174)
• (Contrib) processor/transformprocessor: Add support for flat configuration style. (#29017)
  The flat configuration style allows users to configure statements by providing a list of statements instead of a
  structured configuration map. The statement's context is expressed by adding the context's name prefix to path names,
  which are used to infer and to select the appropriate context for the statement.
• (Contrib) receiver/httpcheck: Added support for specifying multiple endpoints in the httpcheckreceiver using the endpoints field. Users can now monitor multiple URLs with a single configuration block, improving flexibility and reducing redundancy. (#37121)
• (Contrib) processor/resourcedetection: Expose additional configuration parameters for the AWS metadata client used by the EC2 detector (#35936)
  In some cases, you might need to change the behavior of the AWS metadata client from the standard retryer
  By default, the client retries 3 times with a max backoff delay of 20s.
  We offer a limited set of options to override those defaults specifically, such that you can set the client to retry 10 times, for up to 5 minutes, for example:

  processors:
    resourcedetection/ec2:
      detectors: ["ec2"]
      ec2:
        max_attempts: 10
        max_backoff: 5m

• (Contrib) processor/cumulativetodelta: Add metric type filter for cumulativetodelta processor (#33673)
• (Contrib) processor/resourcedetection: Add fail_on_missing_metadata option on EC2 detector (#35936)
  If the EC2 metadata endpoint is unavailable, the EC2 detector by default ignores the error.
  By setting fail_on_missing_metadata to true on the detector, the user will now trigger an error explicitly,
  which will stop the collector from starting.
• (Contrib) processor/resourcedetection: The gcp resource detector will now detect resource attributes identifying a GCE instance's managed instance group. (#36142)
• (Contrib) receiver/jaeger: Log the endpoints of different servers started by jaegerreceiver (#36961)
  This change logs the endpoints of different servers started by jaegerreceiver. It simplifies debugging by ensuring log messages match configuration settings.
• (Contrib) receiver/hostmetrics/process: Added support for tracking process.uptime (#36667)
• (Contrib) receiver/googlecloudpubsub: Added support for encoding extensions. (#37109)
• (Contrib) processor/transform: Replace parser collection implementations with ottl.ParserCollection and add initial support for expressing statement's context via path names. (#29017)
• (Contrib) receiver/prometheus: Add receiver.prometheusreceiver.UseCollectorStartTimeFallback featuregate for the start time metric adjuster to use the collector start time as an approximation of process start time as a fallback. (#36364)
• (Contrib) processor/tailsampling: Reworked the consume traces, sampling decision, and policy loading paths to improve performance and readability (#37560)

🧰 Bug fixes 🧰

• (Core) exporters: Fix bug that the exporter with new batcher may have been marked as non mutation. (#12239)
  Only affects users that manually turned on exporter.UsePullingBasedExporterQueueBatcher featuregate.
• (Core) exporters: Fix MergeSplit issue that ignores the initial message size. (#12257)
• (Core) service-telemetry: pass the missing async error channel into service telemetry settings (#11417)
• (Contrib) receiver/filelog: Fix issue where flushed tokens could be truncated. (#35042)
• (Contrib) connector/routing: Fix config validation with context other than resource (#37410)
• (Contrib) processor/k8sattributes: Wait for the other informers to complete their initial sync before starting the pod informers (#37056)
• (Contrib) processor/metricsgeneration: Generated metric name may not match metric being scaled (#37474)
• (Contrib) connector/routing: The connector splits the original payload so that it may be emitted in parts to each route. (#37390)
• (Contrib) pkg/stanza: Fix default source identifier in recombine operator (#37210)
  Its default value is now aligned with the semantic conventions: attributes["log.file.path"]
• (Contrib) processor/tailsampling: Fixed sampling decision metrics otelcol_processor_tail_sampling_sampling_trace_dropped_too_early and otelcol_processor_tail_sampling_sampling_policy_evaluation_error_total, these were sometimes overcounted. (#37212)
  As a result of this change non-zero values of otelcol_processor_tail_sampling_sampling_trace_dropped_too_early
  and otelcol_processor_tail_sampling_sampling_policy_evaluation_error_total metrics will be lower.
  Before this fix, errors got counted several times depending on the amount of traces being processed
  that tick and where in the batch the error happened.
  Zero values are unaffected.
• (Contrib) exporter/signalfx: Warn on dropping metric data points when they have more than allowed dimension count (#37484)
  The SignalFx exporter drops metric data points if they have more than 36 dimensions.
  Currently, the exporter logs at debug level when this occurs.
  With this change, the exporter will log at the warning level.

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.119.0
  • digest: sha256:2cd525f1e24a461a08beb2f853181c817e70ed53e92932d3bdac213f2449f202
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.119.0
  • digest: sha256:8bc28dd946efc310b67aabf5338a3b5aa89d10a1b23d6f22a5fcc04a1117eb8f

v0.118.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.118.0 and the opentelemetry-collector-contrib v0.118.0 releases where appropriate.

🛑 Breaking changes 🛑

• (Splunk) Stop supporting deprecated syntax for config source expansion (#5832)
  Use the following guidelines to update your configuration:
  • $ENV must be replaced with ${env:ENV}
  • $include:file_path must be replaced with ${include:file_path}. The same applied for any other config source.
    More information can be found in (the upgrade guidelines).
• (Contrib) pkg/stanza: Move filelog.container.removeOriginalTimeField feature gate to stable (#33389)
• (Contrib) pkg/ottl: Support dynamic indexing of maps and slices. (#36644)
• (Contrib) routingconnector: Disconnect match_once parameter from functionality. (#29882)
  The parameter will be ignored, except to trigger a warning log about its upcoming removal in v0.120.0.

💡 Enhancements 💡

• (Splunk) Turn on splunk.continuousDiscovery feature gate by default (#5833)
  This enables new way to discover third-party components and collect metrics from them:
  • It allows discovering services that were not available at the time of the collector startup but were started later.
  • Information about discovered services will be sent to Splunk Observability Cloud that can be found in "Discovered services" page.
    If you want to disable this feature and bring back the old behavior, add --feature-gates=-splunk.continuousDiscovery command line flag.
• (Contrib) mysqlreceiver: Add delete_multi, update_multi metrics to mysql.commands for mysqlreceiver (#37301)
• (Contrib) pkg/ottl: Add the Nanosecond converter to return the nanosecond component from the specified time.Time (#37042)
• (Contrib) pkg/ottl: Add the Second converter to return the second component from the specified time.Time (#37042)
• (Contrib) vcenterreceiver: Adds vCenter memory capacity for hosts and memory granted metric for VMs. (#37257)
• (Contrib) awss3exporter: Implement sending queue for S3 exporter (#37274, #36264)
• (Contrib) resourcedetectionprocessor: This enhancement detects AWS EKS cloud account ID (#37179)
• (Contrib) k8sattributesprocessor: For pods with only one container, the container.id and k8s.container.name are not longer required in the resource attributes to add the container attributes (#34189)
• (Contrib) pkg/ottl: Enhanced error messages for invalid cache access and introduced options to configure their values within the OTTL contexts. (#29017)
• (Contrib) pkg/ottl: Add the FormatTime function to convert time.Time values to human-readable strings (#36870)
• (Contrib) resourcedetectionprocessor: Introduce kubeadm detector to retrieve local cluster name. (#35116)
• (Contrib) signalfxexporter: Prioritize retrieving token from context when accesstokenpassthrough is enabled (#37102)
• (Core) exporterhelper: Add blocking option to control queue behavior when full (#12090)
• (Core) debugexporter: Add EventName to debug exporter for Logs. EventName was added as top-level field in the LogRecord from 1.5.0 of proto definition. (#11966)
• (Core) confighttp: Added support for configuring compression levels. (#10467)
  A new configuration option called CompressionParams has been added to confighttp. | This allows users to configure the compression levels for the confighttp client.
• (Core) exporterhelper: Change the memory queue implementation to not pre-allocate capacity objects. (#12070)
  This change improves memory usage of the collector under low utilization and is a prerequisite for supporting different other size limitations (number of items, bytes).

🧰 Bug fixes 🧰

• (Contrib) googlecloudpubsubreceiver: Fix a goroutine leak during shutdown. (#30438)
  A goroutine leak was found in the googlecloudpubsubreceiver.
  The goroutine leak was caused by the receiver not closing the underlying created gRPC client when using an insecure custom endpoint.
• (Contrib) signalfxexporter: Honor access_token_passthrough config option for sending events (#37102)
• (Contrib) k8sattributesprocessor: Ensure the pods gathered by the processor contain the information about their related replica sets and deployments after the initial sync (#37056)
• (Contrib) pkg/ottl: Fix bug with replace_all_matches and replace_all_patterns that caused non-string values to be changed to empty string when matching against empty string. (#37071)
• (Contrib) tailsamplingprocessor: Fixed sampling policy evaluation debug logging batch metrics (e.g. sampled). (#37040)
• (Contrib) tailsamplingprocessor: Late span age histogram should include sampled traces. (#37180)

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.118.0
  • digest: sha256:76eda47c813540a5ac4ab70abbd0e780a23f11f9fd308d2ec5a9c113529ce54f
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.118.0
  • digest: sha256:ac1de2a46c8108845ebefb90b48714387bd6237de0ed10b3e3765eeeb516a3c4

v0.117.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.117.0 and the opentelemetry-collector-contrib v0.117.0 releases where appropriate.

⚠️ Known issues ⚠️

The sh executable that ships as part of the Docker image of the collector is corrupted in this release.

🛑 Breaking changes 🛑

• (Contrib) cloudfoundryreceiver: Introduce a feature gate enable copying envelope tags to the metrics as resource attributes instead of datapoint attributes. (#34824)
• (Contrib) pkg/ottl: removed the ability to reference entire parent objects. (#36872)
  Statements like set(cache["resource"], resource) in non-resource contexts will no longer work.
• (Contrib) routingconnector: Change default value of match_once parameter to true. (#29882)
  This field was deprecated in v0.116.0 and will be removed in v0.120.0.
• (Core) otelcol: Remove warnings when 0.0.0.0 is used (#11713, #8510)

🚩 Deprecations 🚩

• (Contrib) sapmexporter: Deprecate SAPM exporter (#36028)
  The SAPM exporter is being marked as deprecated. Please use the otlphttp exporter with the configuration shown
  below. Also update your pipeline configuration for Traces accordingly.

    exporters:
        otlphttp:
            traces_endpoint: "${SPLUNK_INGEST_URL}/v2/trace/otlp"
            headers:
                "X-SF-Token": "${SPLUNK_ACCESS_TOKEN}"

🚀 New components 🚀

• (Splunk) Add metricsgeneration processor (#5769)

💡 Enhancements 💡

• (Splunk) Add a new discovery bundle for Envoy proxy metrics (#5780)
• (Contrib) k8sclusterreceiver: Add additional attributes to node and pod entities (#35879)
  Adds the following attributes to node and pod metadata/entities:
  • k8s.pod.phase: The phase of a Pod indicates where the Pod is in its lifecycle. E.g. 'Pending', 'Running'
  • k8s.pod.status_reason: A brief message indicating details about why the pod is in this state. E.g. 'Evicted'
  • k8s.node.condition_*: The condition of a node. e.g. k8s.node.condition_ready. The value can be true, false, unknown.
• (Contrib) receivercreator: Add support for starting logs' collection based on provided k8s annotations' hints (#34427)
• (Contrib) ottl: Add a new ottl trim function that trims leading and trailing characters from a string (default- whitespace). (#34100)
• (Contrib) sqlqueryreceiver: Add instrumentation scope to SQL query receiver metrics and logs (#31028)
• (Contrib) statsdreceiver: Add UDS support to statsdreceiver (#21385)
• (Contrib) tailsamplingprocessor: Support hot sampling policy loading (#37014)

🧰 Bug fixes 🧰

• (Splunk) Fix deprecation warning for multiline config source calls (#5829)
• (Contrib) receiver/azureeventhub: Ensure that observed timestamp is set when unmarshaling logs. (#36861)
• (Contrib) internal/docker: Fix image matching regular expression to properly match SHA256 strings. (#36239)
  This affects the docker_observer extension.
• (Contrib) k8sobjectsreceiver: ensure the k8s.namespace.name attribute is set for objects retrieved using the watch mode (#36352)
• (Contrib) mongodbatlasreceiver: Update the mongoDB Atlas receiver to use the Default HTTP Transport that supports default proxy configuration (#36412)
• (Contrib) mysqlreceiver: Avoid recording a value for the MysqlBufferPoolPages metric when out-of-bounds. (#35495)
  When using compressed tables, Innodb_buffer_pool_pages_misc may report an out-of-bounds value.
  See https://bugs.mysql.com/bug.php?id=59550 for context.
• (Contrib) pkg/ottl: fix handling of nested maps within slices in the flatten function (#36162)
• (Contrib) mysqlreceiver: Divide large values directly in SQL queries to avoid int overflows (#35495)
• (Core) internal/sharedcomponent: Fixed bug where sharedcomponent would use too much memory remembering all the previously reported statuses (#11826)

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.117.0
  • digest: sha256:c07c949d244685030ad8ccd96bc8137b3d229a81e9dafc27c4f2e349e42d3cbb
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.117.0
  • digest: sha256:5b3df3fe83340b7f469d37cecd9ecd92805e97b09e28810547ddddfbf651156c

v0.116.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.116.0 and the opentelemetry-collector-contrib v0.116.0 releases where appropriate.

🛑 Breaking changes 🛑

• (Contrib) processor/tailsampling: Reverts #33671, allowing for composite policies to specify inverted clauses in conjunction with other policies. This is a change bringing the previous state into place, breaking users who rely on what was introduced as part of #33671. (#34085)

🚀 New components 🚀

• (Splunk) Add bearertokenauth extension (#5727)
• (Splunk) Add snowflake receiver (#5724)
• (Splunk) Add purefa receiver (#5731)
• (Splunk) Add sum connector (#5730 )

🚩 Deprecations 🚩

• (Contrib) connector/routing: Deprecate match_once parameter. (#29882)
• (Contrib) routingprocessor: Deprecated in favor of the routing connector. (#36616)
• (Contrib) splunkhecexporter: Add otel_attrs_to_hec_metadata/* config fields to replace hec_metadata_to_otel_attrs/* fields. (#35092)
  otel_attrs_to_hec_metadata/* config fields will replace the hec_metadata_to_otel_attrs/* fields in a later release.

💡 Enhancements 💡

• (Contrib) purefareceiver: Implements support for scraping Pure Storage FlashArray with Purity version 6.6.11+ (#36251)
• (Contrib) azureeventhubreceiver: support providing one or more time formats for timestamp parsing (#36650)
• (Contrib) fileexporter: Add support for profiles signal. (#35978)
• (Contrib) pkg/ottl: GetXML Converter now supports selecting text, CDATA, and attribute (value) content. (#36821)
• (Contrib) kafkaexporter, kafkareceiver: Add a new mechanism "AWS_MSK_IAM_OAUTHBEARER" for kafka exporter and kafka receiver. This mechanism use the AWS MSK IAM SASL Signer for Go https://github.com/aws/aws-msk-iam-sasl-signer-go. (#19747)
• (Contrib) loadbalancingexporter: Adds a an optional configuration to the k8s resolver which returns hostnames instead of IPs for headless services pointing at statefulsets (#18412)
• (Contrib) mongodbatlasreceiver: Adds additional metrics to the MongoDB Atlas receiver (#36525)
  Adds a number of new default disabled metrics to the MongoDB Atlas receiver. These metrics are:
  • mongodbatlas.disk.partition.queue.depth
  • mongodbatlas.disk.partition.throughput
  • mongodbatlas.process.cache.ratio
• (Contrib) s3exporter: Upgrading to adopt aws sdk v2 (#36699)

🧰 Bug fixes 🧰

• (Core) exporterhelper: Fix memory leak at exporter shutdown (#11401)
• (Core) sharedcomponent: Remove race-condition and cleanup locking (#11819)
• (Contrib) receiver/prometheusreceiver: Metric adjuster no longer assumes that all metrics from a scrape come from the same resource (#36477)
• (Contrib) k8sattributesprocessor: Override extracted k8s attributes if original value has been empty (#36373)
• (Contrib) k8sattributesprocessor: Log any errors encountered during kube client initialisation (#35879)
  This addresses an issue where the collector, due to an error encountered during the kubernetes client initialisation,
  was reporting an 'unavailable' status via the health check extension without any further information to be found in the logs.
• (Contrib) postgresqlreceiver: Update the postgresqlreceiver to handle new table schema for the bgwriter metrics in pg17+ (#36784)
• (Contrib) jaegerreceiver: Remove ineffectual warning from jaegerreceiver (#35894)

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.116.0
  • digest: sha256:375ed345d09d7b6c984e002a70f5f8f9881ffd250ca10fb519ac506e8a229bf0
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.116.0
  • digest: sha256:3aa04423420c06868802c55c0a113a9e749500d377b821e2552cc5aad6f6075c

v0.115.0

🛑 Breaking changes 🛑

• (Splunk) Change default traces exporter to otlphttp exporter. (#5588).
  See the upgrade guidelines on how to change existing configs to replace SAPM exporter.
• (Core) Change all logged timestamps to ISO8601 (#10543).
  This makes log timestamps human-readable (as opposed to epoch seconds in
  scientific notation), but may break users trying to parse logged lines in the
  old format.
• (Contrib) k8sattributesprocessor: Move k8sattr.fieldExtractConfigRegex.disallow feature gate to Beta. (#25128)
  Disable the k8sattr.fieldExtractConfigRegex.disallow feature gate to get the old behavior.

🚀 New components 🚀

• (Splunk) Add filestats receiver (#5229)
• (Splunk) Add iis receiver (#5717)

💡 Enhancements 💡

• (Splunk) Automatic Discovery:
  • Switch bundled NGINX discovery to create OpenTelemetry NGINX receiver instead of the Smart Agent NGINX monitor (#5689)
• (Splunk) Expose internal metrics at default localhost:8888 address instead of ${SPLUNK_LISTEN_INTERFACE}:8888 (#5706)
  This can be changed in service::telemetry::metrics section:

  service:
    telemetry:
      metrics:
        readers:
          - pull:
              exporter:
                prometheus:
                  host: localhost
                  port: 8888

  This also removes a warning about deprecated service::telemetry::metrics::address.
• (Core) exporterqueue: Introduce a feature gate exporter.UsePullingBasedExporterQueueBatcher to use the new pulling model in exporter queue batching. (#8122, #10368)
  If both queuing and batching is enabled for exporter, we now use a pulling model instead of a
  pushing model. num_consumer in queue configuration is now used to specify the maximum number of
  concurrent workers that are sending out the request.
• (Core) service: Add support for ca certificates in telemetry metrics otlp grpc exporter (#11633)
  Before this change the Certificate value in config was silently ignored.
• (Contrib) postgresqlreceiver: Added new postgresql metrics to achieve parity with Telegraf (#36528)
• (Contrib) loadbalancingexporter: Adding sending_queue, retry_on_failure and timeout settings to loadbalancing exporter configuration (#35378)
  When switching to top-level sending_queue configuration - users should carefully review queue size
  In some rare cases setting top-level queue size to n*queueSize might be not enough to prevent data loss
• (Contrib) routingconnector: Add abiilty to route by 'datapoint' context (#36523)
• (Contrib) signalfxreceiver: Follow receiver contract based on type of error (#5909)
  Use 503 error code for retryable and 400 error code for not-retryable errors instead of responding with a 500 unconditionally.

🧰 Bug fixes 🧰

• (Core) service: ensure OTLP emitted logs respect severity (#11718)
• (Core) featuregate: Fix an unfriendly display message runtime error when feature gate is used to display command line usage. (#11651)
• (Contrib) exporter/pulsarexporter: Change configuration option map_connections_per_broker, rename to max_connections_per_broker. (#36579)
• (Contrib) processor/k8sattribute: fixes parsing of k8s image names to support images with tags and digests. (#36131)
• (Contrib) loadbalancingexporter: The k8sresolver in loadbalancingexporter was triggering exporter churn in the way the change event was handled. (#35658)
• (Contrib) vcenterreceiver: The existing code did not honor TLS settings beyond 'insecure'. All TLS client config should now be honored. (#36482)

🚩Deprecations 🚩

• (Splunk) Deprecate the collectd/genericjmx monitor. Please use the jmxreceiver instead. (#5539)
• (Splunk) Deprecate the collectd/activemq monitor. Please use the jmxreceiver with the activemq target system instead. (#5539)
• (Splunk) Deprecate the collectd/cassandra monitor. Please use the jmxreceiver with the cassandra target system instead. (#5539)
• (Splunk) Deprecate the collectd/hadoop monitor. Please use the jmxreceiver with the hadoop target system instead. (#5539)
• (Splunk) Deprecate the collectd/kafka monitor. Please use the jmxreceiver with the kafka target system instead. (#5539)
• (Splunk) Deprecate the collectd/kafka-consumer monitor. Please use the jmxreceiver with the kafka-consumer target system instead. (#5539)
• (Splunk) Deprecate the collectd/kafka-producer monitor. Please use the jmxreceiver with the kafka-producer target system instead. (#5539)
• (Splunk) Deprecate the collectd/solr monitor. Please use the jmxreceiver with the solr target system instead. (#5539)
• (Splunk) Deprecate the collectd/tomcat monitor. Please use the jmxreceiver with the tomcat target system instead. (#5539)

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.115.0
  • digest: sha256:1f7f859a1ccccea22c1d1a32bb2f807b8368660b386851be13db02d706ea3b7c
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.115.0
  • digest: sha256:f03d72285f53bc14f88e71a3e323713b752dea5a0581eeddc285f44e74418215

v0.114.0

💡 Enhancements 💡

• (Contrib) processor/k8sattributes: Add support for profiles signal (#35983)
• (Contrib) receiver/k8scluster: Add support for limiting observed resources to a specific namespace. (#9401)
  This change allows to make use of this receiver with Roles/RoleBindings, as opposed to giving the collector cluster-wide read access.
• (Contrib) processor/resourcedetection: Introduce support for Profiles signal type. (#35980)
• (Contrib) connector/routing: Add ability to route by metric context (#36236)
• (Contrib) connector/routing: Add ability to route by span context (#36276)
• (Contrib) processor/spanprocessor: Add a new configuration option to keep the original span name when extracting attributes from the span name. (#36120)
• (Contrib) receiver/splunkenterprise: Add new metrics for Splunk Enterprise dispatch artifacts caches (#36181)

🚩Deprecations 🚩

• (Splunk) SPLUNK_TRACE_URL environment variable is deprecated. It's replaced with ${SPLUNK_INGEST_URL}/v2/trace
  in the default configs. Default value for SPLUNK_TRACE_URL is still set in the binary from SPLUNK_REALM or
  SPLUNK_INGEST_URL environment variables to not break existing configurations. However, it is recommended to
  update the configurations to use ${SPLUNK_INGEST_URL}/v2/trace instead. (#5672).

🛑 Breaking changes 🛑

• (Splunk) Given that SPLUNK_TRACE_URL environment variable is deprecated and replaced with
  ${SPLUNK_INGEST_URL}/v2/trace in the default configurations, the option to set the Trace URL has been removed from
  all packaging and mass deployment solutions to an avoid confusion. (#5672).

🧰 Bug fixes 🧰

• (Splunk) receiver/journald: Upgrade journald client libraries in the Collector docker image by taking them from latest Debian image.
  This fixes journald receiver on kubernetes nodes with recent versions of systemd (#5664).
• (Core) scraperhelper: If the scraper shuts down, do not scrape first. (#11632)
  When the scraper is shutting down, it currently will scrape at least once. With this change, upon receiving a shutdown order, the receiver's scraperhelper will exit immediately.
• (Contrib) pkg/stanza: Ensure that time parsing happens before entry is sent to downstream operators (#36213)
• (Contrib) processor/k8sattributes: Block when starting until the metadata have been synced, to fix that some data couldn't be associated with metadata when the agent was just started. (#32556)
• (Contrib) exporter/loadbalancing: Shutdown exporters during collector shutdown. This fixes a memory leak. (#36024)
• (Contrib) pkg/ottl: Respect the depth option when flattening slices using flatten (#36161)
  The depth option is also now required to be at least 1.
• (Contrib) pkg/stanza: Synchronous handling of entries passed from the log emitter to the receiver adapter (#35453)
• (Contrib) receiver/prometheus: Fix prometheus receiver to support static scrape config with Target Allocator (#36062)

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.114.0
  • digest: sha256:bb2a8cb7e0ba9567f4f3af2be57787220ff808db258b53bd6bcb7d691efda93b
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.114.0
  • digest: sha256:866331331dfdc8ee6ac53537a5211f2c82891b0388fe944001d7e6b474efd502

v0.113.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.113.0 and the opentelemetry-collector-contrib v0.113.0 releases where appropriate.

🛑 Breaking changes 🛑

• (Contrib) sapmreceiver: Remove the deprecated access_token_passthrough from SAPM receiver. (#35972)
  Please use include_metadata instead with the following config option applied to the batch processor:

  batch:
    metadata_keys: [X-Sf-Token]

• (Contrib) pkg/ottl: Promote processor.transform.ConvertBetweenSumAndGaugeMetricContext feature gate to Stable (#36216)
  This gate can no longer be disabled. The convert_sum_to_gauge and convert_gauge_to_sum may now only be used with the metric context.

💡 Enhancements 💡

• (Contrib) splunkenterprisereceiver: Add telemetry around the Splunk Enterprise kv-store. (#35445)
• (Contrib) journaldreceiver: adds ability to parse journald's MESSAGE field as a string if desired (#36005)
• (Contrib) journaldreceiver: allows querying a journald namespace (#36031)
• (Contrib) hostmetricsreceiver: Add the system.uptime metric in the hostmetrics receiver (#31627)
  This metric is provided by the new system scraper.
• (Contrib) hostmetrics: Adjust scraper creation to make it so the scraper name is reported with hostmetrics scraper errors. (#35814)
• (Contrib) pkg/ottl: Add SliceToMap function (#35256)
• (Contrib) journaldreceiver: Restart journalctl if it exits unexpectedly (#35635)
• (Contrib) routingconnector: Add ability to route by request metadata. (#19738)
• (Contrib) exporter/signalfx: Enabling retrying for dimension properties update without tags in case of 400 response error. (#36044)
  Property and tag updates are done using the same API call. After this change, the exporter will retry once to sync
  properties in case of 400 response error.
• (Contrib) signalfxexporter: Add more default metrics related to Kubernetes cronjobs, jobs, statefulset, and hpa (#36026)
• (Contrib) simpleprometheusreceiver: Support to set job_name in config (#31502)
• (Contrib) solacereceiver: Add support to the Solace Receiver to convert the new Move to Dead Message Queue and new Delete spans generated by Solace Event Broker to OTLP. (#36071)
• (Contrib) routingconnector: Add ability to route log records individually using OTTL log record context. (#35939)
• (Contrib) splunkenterprisereceiver: Add new metrics for Splunk Enterprise dispatch artifacts (#35950)
• (Core) batchprocessor: Move single shard batcher creation to the constructor (#11594)
• (Core) service: add support for using the otelzap bridge and emit logs using the OTel Go SDK (#10544)

🧰 Bug fixes 🧰

• (Contrib) receiver/windowseventlog: Fix panic when rendering long event messages. (#36179)
• (Contrib) hostmetricsreceiver: Do not set the default value of HOST_PROC_MOUNTINFO to respect root_path (#35990)
• (Contrib) prometheusexporter: Fixes an issue where the prometheus exporter would not shut down the server when the collector was stopped. (#35464)
• (Contrib) k8sobserver: Enable observation of ingress objects if the ObserveIngresses config option is set to true (#35324)
• (Contrib) pkg/stanza: Fixed bug causing Operators with DropOnErrorQuiet to send log entries to the next operator. (#35010)
  This issue was introduced by a bug fix meant to ensure Silent Operators are not logging errors (#35010). With this fix,
  this side effect bug has been resolved.
• (Contrib) splunkhecreceiver: Avoid a memory leak by changing how we record obsreports for logs and metrics. (#35294)
• (Contrib) receiver/filelog: fix record counting with header (#35869)
• (Contrib) connector/routing: Fix detection of duplicate conditions in routing table. (#35962)
• (Contrib) solacereceiver: The Solace receiver may unexpectedly terminate on reporting traces when used with a memory limiter processor and under high load (#35958)
• (Contrib) pkg/stanza/operator: Retain Operator should propagate the severity field (#35832)
  The retain operator should propagate the severity field like it does with timestamps.
• (Contrib) pkg/stanza: Handle error of callback function of ParserOperator.ProcessWithCallback (#35769)
  ProcessWithCallback of ParserOperator first calls the ParseWith method
  which properly handles errors with HandleEntryError.
  Then the callback function is called and its returned error should also
  be handled by the HandleEntryError ensuring a consistent experience.
• (Core) service: ensure traces and logs emitted by the otel go SDK use the same resource information (#11578)
• (Core) config/configgrpc: Patch for bug in the grpc-go NewClient that makes the way the hostname is resolved incompatible with the way proxy setting are applied. (#11537)

Docker Image Manifests:

• Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector:0.113.0
  • digest: sha256:d5758bd96efd837be26c2351f740a9e7723241bdccb6374b14d18da0a274fbd1
• Windows (2019 amd64, 2022 amd64):
  • quay.io/signalfx/splunk-otel-collector-windows:0.113.0
  • digest: sha256:f4ae66b62f79c75a56fd731de672003efcd50323b7fc02526f16461601567e5f

v0.112.0

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.112.0 and the opentelemetry-collector-contrib v0.112.0 releases where appropriate.

🛑 Breaking changes 🛑

• (Splunk) Remove httpsink exporter (#5503)
• (Splunk) Remove signalfx-metadata and collectd/metadata monitors (#5508)
  Both monitors are deprecated and replaced by the hostmetricsreceiver and processlist monitor.
• (Splunk) Remove deprecated collectd/etcd monitor. Please use the etcd prometheus endpoint to scrape metrics. (#5520)
• (Splunk) Remove deprecated collectd/health-checker monitor. (#5522)
• (Splunk) Remove deprecated loggingexporter from the distribution (#5551)
• (Core) service: Remove stable gate component.UseLocalHostAsDefaultHost (#11412)

🚩Deprecations 🚩

• (Splunk) Deprecate cloudfoundry monitor (#5495)
• (Splunk) Deprecate the heroku observer. Use the resource detection observer with heroku detector instead. (#5496)
• (Splunk) Deprecate mongodb atlas monitor. Please use the mongodbatlasreceiver instead (#5500)
• (Splunk) Deprecate python-monitor monitor (#5501)
• (Splunk) Deprecate windowslegacy monitor (#5518)
• (Splunk) Deprecate statsd monitor. Use the statsd receiver instead. (#5513)
• (Splunk) Deprecate the collectd/consul monitor. Please use the statsd or prometheus receiver instead. See https://developer.hashicorp.com/consul/docs/agent/monitor/telemetry for more information. (#5521)
• (Splunk) Deprecate collectd/mysql monitor. Use the mysql receiver instead. (#5538)
• (Splunk) Deprecate the collectd/nginx monitor. Please use the nginx receiver instead. (#5537)
• (Splunk) Deprecate the collectd/chrony monitor. Please use the chronyreceiver instead. (#5536)
• (Splunk) Deprecate the collectd/statsd monitor. Please use the statsdreceiver instead. (#5542)
• (Splunk) Deprecate the ecs-metadata monitor (#5541)
• (Splunk) Deprecate the collectd/statsd monitor. Please use the statsdreceiver instead. (#)
• (Splunk) Deprecate the haproxy monitor. Please use the haproxyreceiver instead. (#5543)
• (Contrib) sapmreceiver: Deprecate SAPM receiver (#32125)

🚀 New components 🚀

• (Splunk) Add chrony receiver (#5535)

💡 Enhancements 💡

• (Splunk) Update Python to 3.13.0 (5552)
• (Core) confighttp: Adding support for lz4 compression into the project (#9128)
• (Core) service: Hide profiles support behind a feature gate while it remains alpha. (#11477)
• (Core) exporterhelper: Retry sender will fail fast when the context timeout is shorter than the next retry interval. (#11183)
• (Contrib) azureeventshubreceiver: Updates the Azure Event Hub receiver to use the new Resource Logs translator. (#35357)
• (Contrib) pkg/ottl: Add ConvertAttributesToElementsXML Converter (#35328)
• (Contrib) azureblobreceiver: adds support for using azidentity default auth, enabling the use of Azure Managed Identities, e.g. Workload Identities on AKS (#35636)
  This change allows to use authentication type "default", which makes the receiver use azidentity default Credentials,
  which automatically picks up, identities assigned to e.g. a container or a VirtualMachine
• (Contrib) k8sobserver: Emit endpoint per Pod's container (#35491)
• (Contrib) mongodbreceiver: Add support for MongoDB direct connection (#35427)
• (Contrib) chronyreceiver: Move chronyreceiver to beta (#35913)
• (Contrib) pkg/ottl: Parsing invalid statements and conditions now prints all errors instead of just the first one found. (#35728)
• (Contrib) pkg/ottl: Add ParseSimplifiedXML Converter (#35421)
• (Contrib) routingconnector: Allow routing based on OTTL Conditions (#35731)
  Each route must contain either a statement or a condition.
• (Contrib) sapmreceiver: Respond 503 on non-permanent and 400 on permanent errors (#35300)
• (Contrib) hostmetricsreceiver: Use HOST_PROC_MOUNTINFO as part of configuration instead of environment variable (#35504)
• (Contrib) pkg/ottl: Add ConvertTextToElements Converter (#35364)

🧰 Bug fixes 🧰

• (Core) processorhelper: Fix issue where in/out parameters were not recorded when error was returned from consumer. (#11351)
• (Contrib) metricstransform: The previously removed functionality of aggregating against an empty label set is restored. (#34430)
• (Contrib) filelogreceiver: Supports add_metadata_from_filepath for Windows filepaths (#35558)
• (Contrib) filelogreceiver: Suppress errors on EBADF when unlocking files. (#35706)
  This error is harmless and happens regularly when delete_after_read is set. This is because we acquire the lock right at the start of the ReadToEnd function and then defer the unlock, but that function also performs the delete. So, by the time it returns and the defer runs the file descriptor is no longer valid.
• (Contrib) kafkareceiver: Fixes issue causing kafkareceiver to block during Shutdown(). (#30789)
• (Contrib) hostmetrics receiver: Fix duplicate filesystem metrics (#34635, #34512)
  The hostmetrics exposes duplicate metrics of identical mounts exposed in namespaces. The duplication causes errors in exporters that are sensitive to duplicate metrics. We can safely drop the duplicates as the metrics should be exactly the same.
• (Contrib) pkg/ottl: Allow indexing string slice type (#29441)
• (Contrib) mysqlreceiver: Add replica metric support for versions of MySQL earlier than 8.0.22. (#35217)
• (Contrib) stanza/input/windows: Close remote session while resubscribing (#35577)
• (Contrib) `...