opentelemetry-collector

OpenTelemetry Collector Helm Chart

The helm chart installs OpenTelemetry Collector in kubernetes cluster.

Prerequisites

• Helm 3.0+

Installing the Chart

Add OpenTelemetry Helm repository:

helm repo add open-telemetry https://open-telemetry.github.io/opentelemetry-helm-charts

To install the chart with the release name my-opentelemetry-collector, run the following command:

helm install my-opentelemetry-collector open-telemetry/opentelemetry-collector

Configuration

Default configuration

By default this chart will deploy an OpenTelemetry Collector as daemonset with three pipelines (logs, metrics and traces) and logging exporter enabled by default. Besides daemonset (agent), it can be also installed as standalone deployment. Both modes can be enabled together, in that case logs, metrics and traces will be flowing from agents to standalone collectors.

Example: Install collector as a standalone deployment, and do not run it as an agent.

agentCollector:
  enabled: false
standaloneCollector:
  enabled: true

By default collector has the following receivers enabled:

• metrics: OTLP and prometheus. Prometheus is configured only for scraping collector's own metrics.
• traces: OTLP, zipkin and jaeger (thrift and grpc).
• logs: OTLP (to enable container logs, see Configuration for Kubernetes container logs).

There are two ways to configure collector pipelines, which can be used together as well.

Basic top level configuration

Example: Disable metrics pipeline and send traces to zipkin exporter:

config:
  exporters:
    zipkin:
      endpoint: zipkin-all-in-one:14250
  service:
    pipelines:
      metrics: null
      traces:
        exporters:
          - zipkin

Configuration with agentCollector and standaloneCollector properties

agentCollector and standaloneCollector properties allow to override collector configurations and default parameters applied on the k8s pods.

agentCollector(standaloneCollector).configOverride property allows to provide an extra configuration that will be merged into the default configuration.

Example: Enable host metrics receiver on the agents:

agentCollector:
  configOverride:
    receivers:
      hostmetrics:
        scrapers:
          cpu:
          load:
          memory:
          disk:
    service:
      pipelines:
        metrics:
          receivers: [prometheus, hostmetrics]
  extraEnvs:
  - name: HOST_PROC
    value: /hostfs/proc
  - name: HOST_SYS
    value: /hostfs/sys
  - name: HOST_ETC
    value: /hostfs/etc
  - name: HOST_VAR
    value: /hostfs/var
  - name: HOST_RUN
    value: /hostfs/run
  - name: HOST_DEV
    value: /hostfs/dev
  extraHostPathMounts:
  - name: hostfs
    hostPath: /
    mountPath: /hostfs
    readOnly: true
    mountPropagation: HostToContainer

Configuration for Kubernetes container logs

The collector can be used to collect logs sent to standard output by Kubernetes containers. This feature is disabled by default. It has the following requirements:

• It needs agent collector to be deployed, which means it will not work if only standalone collector is enabled.
• It requires the contrib version of the collector image.

To enable this feature, set the agentCollector.containerLogs.enabled property to true and replace the collector image. Here is an example values.yaml:

agentCollector:
  containerLogs:
    enabled: true

image:
  repository: otel/opentelemetry-collector-contrib

command:
  name: otelcol-contrib

The way this feature works is it adds a filelog receiver on the logs pipeline. This receiver is preconfigured to read the files where Kubernetes container runtime writes all containers' console output to.

⚠️ Warning: Risk of looping the exported logs back into the receiver, causing "log explosion"

The container logs pipeline uses the logging console exporter by default. Paired with the default filelog receiver that receives all containers' console output, it is easy to accidentally feed the exported logs back into the receiver.

Also note that using the --log-level=debug option for the logging exporter causes it to output multiple lines per single received log, which when looped, would amplify the logs exponentially.

To prevent the looping, the default configuration of the receiver excludes logs from the collector's containers.

If you want to include the collector's logs, make sure to replace the logging exporter with an exporter that does not send logs to collector's standard output.

Here's an example values.yaml file that replaces the default logging exporter on the logs pipeline with an otlphttp exporter that sends the container logs to https://example.com:55681 endpoint. It also clears the filelog receiver's exclude property, for collector logs to be included in the pipeline.

agentCollector:
  containerLogs:
    enabled: true

  configOverride:
    exporters:
      otlphttp:
        endpoint: https://example.com:55681
    receivers:
      filelog:
        exclude: []
    service:
      pipelines:
        logs:
          exporters:
            - otlphttp

image:
  repository: otel/opentelemetry-collector-contrib

command:
  name: otelcol-contrib

Other configuration options

The values.yaml file contains information about all other configuration options for this chart.