signalfx/splunk-otel-collector-chart splunk-otel-collector-0.151.0

on GitHub

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.151.0.

🛑 Breaking changes 🛑

• chart: Rename all filelog receiver references to file_log (#2396)
  The filelog alias has been deprecated in favor of file_log in the chart-generated configuration.
  Any Helm values or overrides that reference filelog (for example, *.config.receivers.filelog
  or pipelines that list filelog as a receiver) must be updated to use file_log.
  The chart will fail to be installed or upgraded if the deprecated alias is still referenced.

💡 Enhancements 💡

• agent: Add port option for etcd receivers to support non-standard etcd ports (#2402)
  Set agent.controlPlaneMetrics.etcd.port to override the auto-detected default.
• agent, clusterReceiver, gateway: Add OpAMP extension support to the default agent, clusterReceiver and gateway configuration. (#2406)
  The OpAMP extension is included in the default configuration but disabled by default behind a feature gate.
  Use the feature gate ID splunk.opamp.enabled to enable the OpAMP extension.
  Example:
  For the agent: --set agent.featureGates=+splunk.opamp.enabled
  For the clusterReceiver: --set clusterReceiver.featureGates=+splunk.opamp.enabled
  For the gateway: --set gateway.featureGates=+splunk.opamp.enabled
  Note: a new http_forwarder extension binds to port 4320 and forwards requests to Splunk O11y ingest.
  When running on the gateway, the extension is intended as a forwarding mechanism for sending data to the gateway from the agent and clusterReceiver.
  When running on the agent, the extension is intended as an entry point for receiving OpAMP data from instrumentation agents that support OpAMP, such as the Splunk OpenTelemetry Java agent.
• operator: Bump dotnet to v1.14.0 in helm-charts/splunk-otel-collector/values.yaml (#2377)
• operator: Bump java to v2.27.0 in helm-charts/splunk-otel-collector/values.yaml (#2391)

🧰 Bug fixes 🧰

• agent: Fix prometheus/etcd receiver to respect TLS secret and auto-detect scraping mode (#2402)
  When a TLS secret was configured for etcd metrics, the prometheus/etcd receiver ignored it
  and still scraped port 2381 over plain HTTP. It now correctly uses the configured client
  certificates to scrape via mTLS. This fix applies to the prometheus/etcd receiver only
  (used when featureGates.useControlPlaneMetricsHistogramData is enabled, which is the default).
  Enabling etcd metrics no longer requires a TLS secret for environments that don't need
  mTLS (e.g. kubeadm defaults to HTTP on port 2381). See
  Setting up etcd metrics for details.