jaeger

Jaeger Helm Chart

⚠️ Experimental: This chart is under active development with no stability guarantees. Breaking changes may occur in minor versions.

Jaeger is a distributed tracing system. This chart deploys Jaeger v2 using a unified "All-In-One" architecture built on the OpenTelemetry Collector framework.

Changes in v4.2.0

This release is a refactor designed to simplify operation and configuration.

• Unified Architecture: All functionality (Collector, Query, Ingester) is now provided by a single "All-in-One" deployment.
• Configuration: Storage is configured via the config.extensions.jaeger_storage section using native Jaeger/OTEL config syntax.
• Cassandra Schema: Jaeger v2 handles schema creation internally. The legacy schema job has been removed.
• Service Consolidation: A single Service now exposes all ports (agent, collector, query).
• No provisioned storage: Dependency charts have been removed. The user must deploy them separately and configure connection using the Otel config Syntax https://github.com/jaegertracing/jaeger/blob/main/cmd/jaeger/config.yaml

Architecture

This chart uses the All-In-One deployment model.

• Single Binary: Runs as a Deployment scalable to multiple replicas.
• Stateless: Can connect to external persistent storage (Elasticsearch, Cassandra) for production use.
• Default: Memory storage (ephemeral), suitable for testing.
• Scalable: Increase replica count to scale horizontally

Overriding the Jaeger Version

(chart only supports Jaever v2+)

You can customize the Jaeger image and tag using the following values:

jaeger:
  image:
    repository: jaegertracing/jaeger
    tag: "2.2.0"  # Override the default version

Or globally:

tag: "2.2.0"  # Applies to all components

Installing the Chart

Add the Jaeger Tracing Helm repository:

helm repo add jaegertracing https://jaegertracing.github.io/helm-charts

To install a release named jaeger:

helm install jaeger jaegertracing/jaeger

Configuration

There are several examples of how to configure you can reference at:

https://github.com/jaegertracing/jaeger/tree/main/cmd/jaeger

1. In-Memory (Default)

Ideal for testing. No persistence.

helm install jaeger jaegertracing/jaeger

2. Elasticsearch (Production Recommended)

This chart does not provision an ElasticSearch instance.

To use ElasticSearch storage, you must provide your own ElasticSearch instance and configure Jaeger via the native config syntax:

Elasticsearch example:

values.yaml

userconfig:
  extensions:
    jaeger_query:
      storage:
        traces: primary_store_elasticsearch
        traces_archive: archive_store_elasticsearch
    jaeger_storage:
      backends:
        primary_store_elasticsearch:
          elasticsearch:
            server_urls: ["http://my-elasticsearch-master:9200"]
            auth:
              basic:
                username: elastic
                password: changeme
        archive_store_elasticsearch:
          elasticsearch:
            server_urls: ["http://my-elasticsearch-master:9200"]
            auth:
              basic:
                username: elastic
                password: changeme
  exporters:
    jaeger_storage_exporter:
      trace_storage: primary_store_elasticsearch

Running Maintenance Jobs:

To run Index Cleaner or Rollover jobs, enable them. They require configuration by adjusting storage.elasticsearch values. See that section in values.yaml for more details.

esIndexCleaner:
  enabled: true

storage:
  elasticsearch:
    tls:
      enabled: false
    url: http://elasticsearch-master:9200
    # user: elastic
    # password: changeme

• es-rollover requires elasticsearch to be running BEFORE install so the hook job can run.

3. Cassandra

This chart does not provision a Cassandra cluster.

To use Cassandra storage, you must provide your own Cassandra instance and configure Jaeger via the native config syntax:

values.yaml Example:

storage:
  type: cassandra

config:
  extensions:
    jaeger_storage:
      backends:
        primary_store:
          cassandra:
            connection:
              servers:
                - cassandra-host
              port: 9042
            schema:
              keyspace: jaeger_v1_test

Note: The legacy Cassandra schema job has been removed. Jaeger v2 handles schema creation internally.

4. Spark Dependencies

To run the Spark dependencies job (for dependency links graph) set spark.enabled=true

Below is an example of how to set overrides. Please see the values.yaml for more examples:

The values are populated under storage.elasticsearch and storage.casandra Please see comments in values.yaml for more details.

You must choose the storage type with storage.type=elasticsearch or storage.type=cassandra

spark:
  enabled: true
  extraEnv:
    - name: ES_NODES_WAN_ONLY
      value: "true"

For a full list of supported environment variables, see the Spark Dependencies README.

5. Query UI

To access the ui you can either:

• enable the ingress and fill at least one host
• provide your own ingress
• or port forward kubectl port-forward --namespace {{ .Release.Namespace }} $POD_NAME 16686:16686 --address 0.0.0.0 and access at http://localhost:16686/

jaeger:
  ingress:
    enabled: true
    hosts:
      - <fill a host here>

You can customize the UI by setting uiconfig. Please see the values.yaml for examples or https://github.com/jaegertracing/jaeger/tree/main/cmd/jaeger

Configuring the Collector

The Jaeger v2 configuration is defined in config using OpenTelemetry Collector syntax. You can override pipelines, receivers, and processors there.

config:
  service:
    pipelines:
      traces:
        receivers: [otlp, jaeger, zipkin]
        processors: [batch]
        exporters: [jaeger_storage_exporter]

Ports

The unified Service exposes the following ports:

• Query UI: 16686, 16685 (gRPC)
• OTLP: 4317 (gRPC), 4318 (HTTP)
• Jaeger: 14250 (gRPC), 14268 (HTTP), 6831/6832 (UDP)
• Zipkin: 9411
• Prometheus: 8888 (HTTP), 8889 (HTTP)
• Additional ports may be added via jaeger.service.extraPorts override.