splunk-otel-collector-0.138.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.138.0.

🛑 Breaking changes 🛑

• chart: Remove fluentd sidecar container and related configuration options (#2138)

  The legacy fluentd sidecar container has been removed from the chart. All users should now use the native OpenTelemetry logs collection, which is the default and provides better performance and resource efficiency.

  If you still need to use fluentd for log collection, you can run it as a separate deployment using the official fluentd helm chart and forward logs to the collector using the fluentforward receiver.

  See the upgrade guide for detailed instructions.

splunk-otel-collector-0.137.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.137.0.

🛑 Breaking changes 🛑

• agent: Remove signalfx receiver from the default agent configuration (#2120) This change affects the metrics pipeline only. Data should be sent to the agent in the OTLP format instead of SignalFx. If the signalfx receiver is still needed, the following section can be added to your values.yaml file:

  agent:
    config:
      receivers:
        signalfx:
          endpoint: 0.0.0.0:9943
    ports:
      signalfx:
        containerPort: 9943
        hostPort: 9943
        protocol: TCP
        enabled_for: [metrics]
  

  Please note that the signalfx receiver will eventually be deprecated and removed.

💡 Enhancements 💡

• agent: Bind the OTLP HTTP port 4318 to the host network (#2102)
  This change is effective only for Windows nodes, where the port binding was previously missing.
  For Linux nodes, the hostNetwork setting already ensures the port is bound, so no change in behavior should occur for default chart config.
• agent: Configuration changes to support Secure Application features in Splunk Observability Cloud (#2106)
• chart: Move deprecated Istio pod annotation to pod label (#2108)
  The sidecar.istio.io/inject annotation
  has been deprecated in favor of the label
  with the same name.
• clusterReceiver: Enable additional attributes (k8s.kubelet.version) for k8s cluster receiver (#2103)
• clusterReceiver: Added k8s.hpa.scaletargetref.{api_version,kind,name} resource attributes, plus transform logic mapping HPA scaleTargetRef to workload (replicaset/statefulset/deployment) names in the Kubernetes cluster receiver configuration (#2126)
• gateway: Replace access_token_passthrough with include_metadata in the gateway configuration. This change follows the removal of access_token_passthrough from the signalfx receiver and maintains the previous functionality of forwarding the access token. (#2122)
• gateway: Configuration changes to support Secure Application features in Splunk Observability Cloud (#2128)
• operator: Use splunk-otel-instrumentation-python image for auto-instrumentation of Python applications with the operator. (#1965)
  • To use a different image for Python auto-instrumentation, set the configuration option instrumentation.python.image.
• operator: Bump java to v2.20.1 in helm-charts/splunk-otel-collector/values.yaml (#2070)

🧰 Bug fixes 🧰

• agent: Fix agent configuration when metrics are disabled for Splunk O11y Cloud (#2083)
  Some functionality should not be configured for the agent when metrics are disabled,
  as it causes the agent to crash on startup.
• chart: Fix creation of validateSecret image name to use correct image configuration values (#2123)

splunk-otel-collector-0.136.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.136.1.

🛑 Breaking changes 🛑

• clusterReceiver: Remove the k8sattributesprocessor from the clusterReceiver logs pipeline to fix incorrect enrichment of Kubernetes object and event data. (#2075)
  • The k8sattributesprocessor was previously misconfigured in the clusterReceiver pipeline, resulting in all Kubernetes objects and events being enriched with pod metadata, even when it was not appropriate.
  • This led to incorrect and non-deterministic k8s.pod.* attributes being added to objects and events, such as enriching non-pod objects with pod fields.
  • After this change:
    • Events from the k8s_events receiver will no longer include container.id, container.image.name, and container.image.tag attributes.
    • Events from the k8sobjects receiver will no longer have any k8s.pod.* or container.* attributes except k8s.namespace.name (when applicable) and k8s.resource.name
  • This restores correct and predictable enrichment, and eliminates misleading data.
  • If your downstream systems or dashboards depend on the removed attributes, please update them accordingly.

💡 Enhancements 💡

• opentelemetry-operator-crds: Bump subchart opentelemetry-operator-crds to 0.0.5. Refer to further instructions for updating CRDs if using option. (#2091)
• operator: Bump operator to 0.95.3 in helm-charts/splunk-otel-collector/Chart.yaml (#2091)

🧰 Bug fixes 🧰

• agent: Skip os-release mount for GKE (#2089)

splunk-otel-collector-0.135.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.135.0.

🛑 Breaking changes 🛑

• all: Update labels in all resources (#2027)
  • Remove "heritage" label from all objects managed by the chart
  • Set all "app.kubernetes.io/managed-by" labels to "Helm"
• agent: Remove splunkObservability.logsEnabled option (#2058)
  Logs cannot be sent directly to Splunk Observability anymore.
  Configure splunkPlatform to send logs to Splunk Platform and enable Log Observer Connect to view the logs in Splunk Observability.
  See the following documentation for more details: https://help.splunk.com/en/splunk-observability-cloud/manage-data/view-splunk-platform-logs/accomplish-logs-pipeline-rules-in-splunk-platform

💡 Enhancements 💡

• chart: Adds configuration option for Splunk Platform metrics sourcetype (#1941)
  • Users can override the metrics sourcetype for Splunk Platform using the splunk.com/metricsSourcetype annotation or the .Values.splunkPlatform.metricsSourcetype Helm value.
  • See advanced-configuration.md for detailed precedence and usage notes.

🧰 Bug fixes 🧰

• log: Update sending_queue batch settings when noDropLogsPipeline is enabled (#2076)

splunk-otel-collector-0.134.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.134.0.

🛑 Breaking changes 🛑

• instrumentation: Reduce Java and .Net auto-instrumentation resource attributes (#2006)
  This change limits the resource attributes added to the telemetry data emitted by Java and .Net instrumentation libraries to the attributes generated by the Container and Environment resource providers.
  To enable additional resource providers, update values.yaml environment variable OTEL_JAVA_ENABLED_RESOURCE_PROVIDERS for Java, or set OTEL_DOTNET_AUTO_{DETECTOR}_RESOURCE_DETECTOR_ENABLED for .Net.
  For more information see: https://github.com/signalfx/splunk-otel-collector-chart/blob/main/docs/auto-instrumentation-install.md#4-Resource-Attributes
• chart: Remove deprecated image.fluentd.initContainer configuration option. (#1995)
  This option was deprecated in version 0.38.0. The same splunk/fluentd-hec image is now used in init containers.
• chart: Remove deprecated extraAttributes.podLabels configuration option. (#1994)
  This option was deprecated in version 0.38.0 in favor of extraAttributes.fromLabels.
  See migration guide: https://github.com/signalfx/splunk-otel-collector-chart/blob/main/UPGRADING.md#0371-to-0380

💡 Enhancements 💡

• clusterReceiver: Enable k8s.container.status.last_terminated_reason resource attribute for container metrics (#1968)
• agent: Update prometheus receiver configuration for CoreDNS in OpenShift to use secure endpoint and service account token (#2009)
  • This aligns with OpenShift's default CoreDNS daemonset setup with kube-rbac-proxy which doesn't expose metrics over HTTP.
• operator: Bump java to v2.19.0 in helm-charts/splunk-otel-collector/values.yaml (#1991)
• Target Allocator: Update version to v0.132.0 (#2019)
  Please review the config targetAllocator.config.collector_namespace
  introduced in this version which is set to the release namespace by default.

🧰 Bug fixes 🧰

• agent: Introducing: logsCollection.containers.multilineConfigs.maxNumOfLinesToCombine.
  This new field allows to specify max number of lines which will be combined in log entry for multiline log configuration.
  In some cases, multiline logs with more than 1000 lines were split into multiple events.
  (#2012)
• chart: Fix per-pod configuration of 'metricsIndex' and 'sourcetype' config for Splunk Platform metrics via annotations (#1977)

splunk-otel-collector-0.132.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.132.0.

🛑 Breaking changes 🛑

• chart: Remove support of deprecated distro and provider root-level values.yaml options. (#1989)
  Use distribution instead of distro and cloudProvider instead of provider.
• chart: Remove support of deprecated splunkRealm, splunkAccessToken, ingestUrl, apiUrl, metricsEnabled, tracesEnabled, and logsEnabled root-level values.yaml options. (#1985, #1987, #1988)
  These parameters were deprecated in favor of the structured splunkObservability configuration.
  Users must migrate to using:
  • splunkObservability.realm instead of splunkRealm
  • splunkObservability.accessToken instead of splunkAccessToken
  • splunkObservability.ingestUrl instead of ingestUrl
  • splunkObservability.apiUrl instead of apiUrl
  • splunkObservability.metricsEnabled instead of metricsEnabled
  • splunkObservability.tracesEnabled instead of tracesEnabled
  • splunkObservability.logsEnabled instead of logsEnabled
• chart: Remove support of deprecated otelAgent, otelCollector, and otelK8sClusterReceiver configuration sections (#1990)
  Use agent instead of otelAgent, gateway instead of otelCollector, and clusterReceiver instead of otelK8sClusterReceiver.
  See migration guide: https://github.com/signalfx/splunk-otel-collector-chart/blob/main/UPGRADING.md#0371-to-0380

🚩 Deprecations 🚩

• agent: Mark splunkObservability.logsEnabled as deprecated. (#1984)
  Logs cannot be sent directly to Splunk Observability anymore.
  Configure splunkPlatform to send logs to Splunk Platform and enable Log Observer Connect to view the logs in Splunk Observability.
  See the following documentation for more details: https://help.splunk.com/en/splunk-observability-cloud/manage-data/view-splunk-platform-logs/accomplish-logs-pipeline-rules-in-splunk-platform

💡 Enhancements 💡

• operator: Drop requirement to install all CRDs except instrumentations. (#1969)
• operator: Upgrade the operator subchart version to 0.92.5 (#1973)

splunk-otel-collector-0.131.1

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.131.2.

🚩 Deprecations 🚩

• chart: Deprecated root-level 'service' config in favor of 'agent.service' and 'gateway.service'. (#1943)
  • The 'service' config is now deprecated and will be completely removed in a future release. The default .Values.service is not set anymore.
  • Please use 'agent.service' and 'gateway.service' for configuring services for the agent and gateway, respectively.
  • When the deprecated 'service' config is present, it completely overrides the component-specific service config for backward compatibility. Make sure to set all fields if passing .Values.service.
  • A warning is shown in NOTES.txt if the deprecated config is used.

splunk-otel-collector-0.131.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.131.0.

🛑 Breaking changes 🛑

• agent, clusterReceiver: Changed the default value of the 'useControlPlaneMetricsHistogramData' and 'enableEKSApiServerMetrics' feature gates. These feature gates now default to true and only affect Kubernetes control plane metrics collection. (#1944)
  • Enabling 'featureGates.useControlPlaneMetricsHistogramData' switches the agent's control plane metrics collection from the legacy smartagent/* receivers to prometheus/* receivers for k8s control plane components (apiserver, controller-manager, scheduler, etcd, kube-proxy, coredns).
    • This change also enables sending histogram metrics from the above-listed k8s components as OTLP histograms to Splunk Observability.
    • To disable this feature, set featureGates.useControlPlaneMetricsHistogramData to false.
  • The 'featureGates.enableEKSApiServerMetrics', now set to true, controls whether EKS API server metrics are collected and sent to Splunk Observability, including OTLP histograms.
    • This change only affects clusterReceiver deployments running in EKS (including vanilla, Fargate, or Auto Mode) clusters.
    • To disable this feature, set featureGates.enableEKSApiServerMetrics to false.
  • Please review your configuration if you relied on the previous default of false for these feature gates.

💡 Enhancements 💡

• operator: Bump java to v2.18.0 in helm-charts/splunk-otel-collector/values.yaml (#1935)

🧰 Bug fixes 🧰

• agent: Remove AKS kubeletstat custom CA configuration change. For more information, see the following link https://github.com/signalfx/splunk-otel-collector-chart/blob/main/docs/advanced-configuration.md#aks-kubeletstats-receiver (#1954)

splunk-otel-collector-0.130.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.130.0.

💡 Enhancements 💡

• operator: Bump dotnet to v1.11.0 in helm-charts/splunk-otel-collector/values.yaml (#1927)
• operator: Bump nodejs to v3.3.0 in helm-charts/splunk-otel-collector/values.yaml (#1926)

🧰 Bug fixes 🧰

• gateway: Fix typo in deployment-gateway.yaml (#1939)

splunk-otel-collector-0.129.0

This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.129.0.

🛑 Breaking changes 🛑

• operator: Operator instrumentation configuration is now defined under instrumentation.spec. This change enables full customization of the instrumentation spec, with all user-provided instrumentation fields merged into the default spec. (#1873)
  • The configuration structure for operator auto-instrumentation has changed. All settings must now be defined under instrumentation.spec.
  • Any configuration directly under instrumentation.* (outside of spec) will cause an error during helm operations.
  • If you use .Values.operator.enabled=true and have overrides for instrumentation settings, please review the upgrade guidelines
• agent: Updates agent config for k8s control plane component's prometheus receivers to filter and keep only a limited set of metrics. (#1887)
  • This change is only relevant for users who have enabled the useControlPlaneMetricsHistogramData featuregate in their Helm chart configuration.
• clusterReceiver: Updates scrape config for EKS apiserver prometheus receiver. The change reflects metrics used in out-of-the-box dashboards in Splunk Observability. (#1893)
  • This change is only relevant for users who have enabled the enableEKSApiServerMetrics featuregate in their Helm chart configuration.

💡 Enhancements 💡

• other: Add support for EKS Auto Mode (Part 1, hostnetwork). (#1899)
• other: Remove EC2 detector and expand EKS detector which, in addition to IMDS, supports fetching attributes through API (Part 2 - POD Identity) . (#1903)
  For more information, see the EKS Auto Mode documentation
• clusterReceiver: Enable k8s events in the clusterReceiver by default when exporting logs is enabled. (#1895)
• operator: Bump java to v2.17.1 in helm-charts/splunk-otel-collector/values.yaml (#1917)
• opentelemetry-operator-crds: Bump subchart opentelemetry-operator-crds to 0.0.3. Refer to further instructions for updating CRDs if using option. (#1914)
• operator: Bump operator to 0.91.0 in helm-charts/splunk-otel-collector/Chart.yaml (#1914)

🧰 Bug fixes 🧰

• agent: Fix Signalfx exporter metadata and add necessary mount for bottlerocket OS (#1898)