This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.113.0.
agent
: Fixes hostmetrics receiver to use the correct mount path of the host's filesystem from inside the container (#1547)agent
: Exclude scraping filesystem metrics from mounts that are not accessible from inside the container to avoid scraping errors. (#1550)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.112.0.
agent
: Fix bug where hostmetrics receiver was failing to scrape the filesystem (#1533)operator
: Fix bug where sometimes Instrumentation opentelemetry.io/v1alpha1 can be installed too early (#1544)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.112.0.
agent, gateway, chart
: update default traces exporter to otlphttp (#1518) If you use the sapm exporter with custom settings, you have two options:- Migrate your sapm settings to the new otlphttp exporter.
- Retain SAPM settings by moving them to your agent.config or gateway.config overrides to ensure they remain effective.
operator
: Bump operator to 0.71.2 in helm-charts/splunk-otel-collector/Chart.yaml (#1511)operator
: Bump java to v2.9.0 in helm-charts/splunk-otel-collector/values.yaml (#1509)operator
: Bump nodejs to v2.14.0 in helm-charts/splunk-otel-collector/values.yaml (#1519)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.111.0.
chart
: Added a note anbout the deprecation of thefluentd
option in the chart (#1460)
chart
: Propagated "sourcetype" to work for metrics and traces (#1376)agent
: The agent is now deployed with a Kubernetes service for receiving telemetry data by default (#1485)operator
: Bump dotnet to v1.7.0 in helm-charts/splunk-otel-collector/values.yaml (#1474)operator
: Bump nodejs to v2.13.0 in helm-charts/splunk-otel-collector/values.yaml (#1470)
agent
: Add k8s.node.name attribute to discovered service entities to fix broken link in the UI. (#1494)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.110.0.
template
: Add default metadata key for token to batch processor (#1467) Add default metadata key for token to batch processor. This will allow the token to be retrieved from the context. When SAPM is deprecated and OTLP used, this will be the normal mode of operation.operator
: Bump operator to 0.56.0 in helm-charts/splunk-otel-collector/Chart.yaml (#1446)operator
: Bump java to v2.8.1 in helm-charts/splunk-otel-collector/values.yaml (#1458)
agent
: use root_path to configure the hostmetricsreceiver, instead of environment variables. (#1462)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.109.0.
operator
: Operator Helm values previously under.Values.operator.instrumentation.spec.*
have been moved to.Values.instrumentation.*
(#1436) If you use custom values under.Values.operator.instrumentation.spec.*
please review the upgrade guidelines
-
agent
: AddedfsyncFlag
configuration to allow users to enable fsync on the filestorage. (#1425) -
agent
: Add a feature gateuseControlPlaneMetricsHistogramData
(#1372) This feature gate allows to gather control plane metrics and send them as histogram data to Observability Cloud. This is an experimental feature under heavy development. -
agent
: Add base configuration to support the new continuous discovery mechanism. (#1455) The new continuous discovery mechanism is disabled by default. To enable it, set the following values in your configuration:agent: discovery: enabled: true featureGates: splunk.continuousDiscovery
-
operator
: Bump nodejs to v2.12.0 in helm-charts/splunk-otel-collector/values.yaml (#1434)
targetAllocator
: Fix the name of the service account token given when featureGates.explicitMountServiceAccountToken is true (#1427)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.105.0.
all
: Offer an experimental feature gate to mount service tokens in specific containers. (#1421) Kubernetes API access tokens are currently granted via mounting them on all containers of the cluster receiver, gateway and daemonset. They are also enabled for the target allocator deployment. This experimental change defines how to mount the service account token on specific containers.
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.105.0.
operator
: Bump java from v1.32.3 to v2.7.0 in helm-charts/splunk-otel-collector/values.yaml (#1349) This is a major upgrade. If you use Java auto-instrumentation please review the upgrade guidelines
agent
: Retry indefinitely on filelog receiver if noDropLogsPipeline feature gate is enabled. (#1410)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.105.0.
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.105.0.
operator
: Bump nodejs to v2.11.0 in helm-charts/splunk-otel-collector/values.yaml (#1352)
agent
: Remove the deprecated OTLP HTTP port 55681 (#1359)
targetAllocator
: Add support for Target Allocator as part of the Helm chart. (#689) Target Allocator is a component of the OpenTelemetry Operator. With this addition, the target allocator is deployed to work in coordination with the daemonset of collectors. It applies a default configuration applying scrape targets per node. By default, the Target Allocator looks for all ServiceMonitor and PodMonitor CRDs across all namespaces. This can be tuned by overriding the service account associated with the Target Allocator.
agent
: Add an experimental feature gate to use exporter batching instead of the batch processor (#1387)all
: Set automountServiceAccountToken to true explicitly for the chart's defined service accounts. (#1390)operator
: Bump java to v1.32.3 in helm-charts/splunk-otel-collector/values.yaml (#1355)
agent
: Remove apparmor pod annotation by enabled default (#1378)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.105.0.
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.104.0.
operator
: Bump nodejs to v2.9.0 in helm-charts/splunk-otel-collector/values.yaml (#1337)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.103.0.
operator
: Bump dotnet to v1.6.0 in helm-charts/splunk-otel-collector/values.yaml (#1327)operator
: Bump java to v1.32.2 in helm-charts/splunk-otel-collector/values.yaml (#1328)
chart
: Updated Security Context Constraints for OpenShift support to fix formatting issues and add support for the operator service account (#1325)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.102.1.
agent
: Add a pod annotation that designates the otel-collector as unconfined for appArmor-protected environments (#1313)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.101.0.
operator
: Bump java to v1.32.1 in helm-charts/splunk-otel-collector/values.yaml (#1300)
operator
: Fix issue where SPLUNK_OTEL_AGENT env var was set before custom operator.instrumentation.spec.env env vars (#1292)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.100.0.
agent
: Add an optionskipInitContainers
to skip init container setting file ACLs when therunAsUser
andrunAsGroup
are provided. This is useful when the user wants to manage the file ACLs themselves. (#1286)operator
: Bump dotnet to v1.5.0 in helm-charts/splunk-otel-collector/values.yaml (#1282)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.99.0.
operator
: Bump java to v1.32.0 in helm-charts/splunk-otel-collector/values.yaml (#1231,#1265)operator
: Bump nodejs to v2.8.0 in helm-charts/splunk-otel-collector/values.yaml (#1269)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.98.0.
agent
: Enableretry_on_failure
for journald receiver (#764) In case of temporary errors the journald receiver should slow down and retry the log delivery instead of dropping it.
clusterReceiver
: Added clusterRole for events.k8s.io, without it k8sobjectsreceiver throws an error on startup (#1238)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.97.0.
chart
: Removed memory_ballast property from all the configs (#1240)operator
: Bump operator to 0.49.1 in helm-charts/splunk-otel-collector/Chart.yaml (#1221)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.96.1.
operator
: Bump certmanager to v1.14.4 in helm-charts/splunk-otel-collector/Chart.yaml (#1205)operator
: Bump java to v1.31.0 in helm-charts/splunk-otel-collector/values.yaml (#1199)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.95.0.
chart
: The request is to add the namespace to the generated templates, for commands like "helm template --namespace mynamespace (...)" (#1011)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.94.0.
operator
: Bump certmanager to v1.14.3 in helm-charts/splunk-otel-collector/Chart.yaml (#1182)operator
: Bump java to v1.30.3 in helm-charts/splunk-otel-collector/values.yaml (#1188)operator
: Bump nodejs to v2.7.1 in helm-charts/splunk-otel-collector/values.yaml (#1180)
clusterReceiver
: Bring back the default translations for kubelet metrics in EKS Fargate (#1174)agent
: Remove a post-delete hook which targeted one a single node for reverting file ACLs. (#1175) The removed hook was intended to undo the ACLs set on log directores when runAsUser and runAsGroup are provided. An initContainer run as root-user updates the permissions of log directories to allow read access to the provided uid/gid. But there is no graceful way to revert these ACLs on each node as part of the chart uninstallation process.
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.93.0.
networkExplorer
: ThenetworkExplorer
option was deprecated in 0.88.0. It is now entirely removed. (#1156)
agent
: Fix GKE Autopilot deployment (#1171)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.93.0.
operator
: Improve Auto-instrumentation Configurations (#1166)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.93.0.
agent
: Change the default sending queue size from 5000 to 1000. (#1157)operator
: Bump dotnet to v1.4.0 in helm-charts/splunk-otel-collector/values.yaml (#1158)operator
: Bump certmanager to v1.14.2 in helm-charts/splunk-otel-collector/Chart.yaml (#1152)β οΈ Known Issue fixed in certmanager v1.14.2- In cert-manager v1.14.1, the CA and SelfSigned issuers issue certificates with SANs set to non-critical even when the subject is empty. It incorrectly copies the critical field from the CSR.
- To avoid this issue, please upgrade directly to version 0.93.1 of this chart when utilizing
certmanager.enabled=true
, thereby bypassing affected versions v0.92.1 and v0.93.0.
chart
: Remove networkExplorer deprecation note that can cause the chart installation to fail (#1162)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.93.0.
internal metrics
: Stop reporting metrics sent by the batch processor (#1147)
operator
: Bump java to v1.30.1 in helm-charts/splunk-otel-collector/values.yaml (#1139)operator
: Bump nodejs to v2.7.0 in helm-charts/splunk-otel-collector/values.yaml (#1143)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.92.0.
operator
: Bump certmanager to v1.14.1 in helm-charts/splunk-otel-collector/Chart.yaml (#1140)
chart
: Fix Helm chart incorrectly handling Kubernetes versions containing a "+" character, causing deployment errors for PodDisruptionBudget in certain environments (#1144)collector
: Fix template function to be able to convert non-integer memory limit values (#1128)operator
: Fix issue where the collector agent exporter endpoint used in operator .NET and Python auto-instrumentation was missing the proper IP address (#1129)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.92.0.
operator
: Use the Splunk Distribution of OpenTelemetry .NET docker image by default when auto-instrumenting with the operator (#1098)other
: Added support for rerouting metrics index with pod/namespace annotations (#1053)chart
: Allows to set the hostNetwork parameter in chart (#1014)operator
: Bump dotnet to v1.3.0 in helm-charts/splunk-otel-collector/values.yaml (#1121)operator
: Bump operator to 0.46.0 in helm-charts/splunk-otel-collector/Chart.yaml (#1116,#1124)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.91.3.
agent
: linux: Adopt discovery mode in agent and provide agent.discovery.properties value mapping (#1108)chart
: Make clusterName optional in EKS and GKE (#1056,#1067)operator
: Bump certmanager to v1.13.3 in helm-charts/splunk-otel-collector/Chart.yaml (#1085)operator
: Bump nodejs to v2.6.1 in helm-charts/splunk-otel-collector/values.yaml (#1094)operator
: Bump operator to 0.44.2 in helm-charts/splunk-otel-collector/Chart.yaml (#1084)
agent
: Change the default directory of the journald receiver (#1110)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.91.0.
networkExplorer
: Remove networkExplorer from helm chart (#1076) Network explorer is no longer part of this helm chart and should be installed separately. See https://github.com/signalfx/splunk-otel-collector-chart/blob/main/docs/advanced-configuration.md#using-opentelemetry-ebpf-helm-chart-with-splunk-opentelemetry-collector-for-kubernetes for more details.
operator
: Bump nodejs to v2.6.0 in helm-charts/splunk-otel-collector/values.yaml (#1080)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.90.0.
agent
: Fix GKE Autopilot deployment (#1071)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.90.0.
operator
: Bump java to v1.30.0 in helm-charts/splunk-otel-collector/values.yaml (#1064)operator
: Bump version of the operator subchart to 0.43.0 (#1049)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.88.0.
networkExplorer
: Deprecate networkExplorer in favor of the upstream OpenTelemetry eBPF helm chart. (#1026)
agent
: Remove the use of themax_connections
configuration key, usemax_idle_conns_per_host
instead. (#1034)operator
: Bump java to v1.29.1 in helm-charts/splunk-otel-collector/values.yaml (#1042)operator
: Bump nodejs to v2.5.1 in helm-charts/splunk-otel-collector/values.yaml (#1040)operator
: Bump operator to 0.42.3 in helm-charts/splunk-otel-collector/Chart.yaml (#1025)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.87.0.
agent, fluentd
: Allow users to overrideenable_stat_watcher
andrefresh_interval
for tail plugin from values.yaml (#982)agent
: Add combineWith field to multiline configuration (#756)operator
: Bump certmanager to v1.13.2 in helm-charts/splunk-otel-collector/Chart.yaml (#1007)operator
: Bump operator to 0.41.0 in helm-charts/splunk-otel-collector/Chart.yaml (#985)
chart
: Remove by default empty allowedFlexVolumes (#981) Removed the allowedFlexVolumes empty field since it does not provide any default additional benefits for the users.
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.86.0.
operator
: use splunk-otel-js as Docker image for node.js auto-instrumentation (#967)
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.86.0.
agent
: Drop time attribute used during parsing the time from log record, so it is not reported as an extra field. (#912)agent
: Change the default logs collection engine (logsEngine
) to the native OpenTelemetry logs collection (otel
) (#934) If you want to keep using Fluentd sidecar for the logs collection, setlogsEngine' to 'fluentd
in your values.yamloperator
: Bump certmanager to v1.13.1 in helm-charts/splunk-otel-collector/Chart.yaml (#941)operator
: Bump operator to 0.39.1 in helm-charts/splunk-otel-collector/Chart.yaml (#940)chart
: Add support for OpenTelemetry CHANGELOG.md generator tool, see chloggen (#923)
networkExplorer
: delete deprecated Pod Security Policy from networkExplorer templates (#896)operator
: Fix Operator helm template issues (#938) Resolves smaller issues in the Helm template related to the of non-default Operator auto-instrumentation values, which could trigger deployment failures
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.85.0.
- Add
component
label to daemonset pods and optional k8s service for agent daemonset which can be enabled withagent.service.enabled
config. #740
- Update Splunk Fluend HEC docker image to v1.3.3 #924
- Add ability to update and track operator auto-instrumentation images #917
- [BREAKING CHANGE] Refactored auto-instrumentation image definition from operator.instrumentation.spec.{library}.image to operator.instrumentation.spec.{library}.repository and operator.instrumentation.spec.{library}.tag. See upgrade guidelines
- Upgrade the Splunk OpenTelemetry Collector for Kubernetes dependencies #32,#929
- Disable
opencensus.resourcetype
resource attribute in k8s_cluster receiver #914- This change does not affect Splunk Observability users since it has already been disabled in the default translation rules of the Signalfx exporter
- Disable signalfx exporter default translations in clusterReceiver deployment #915
- This change improves performance of clusterReceiver, but can be breaking if the deprecated signalfx exporter
translation_rules
option is being used
- This change improves performance of clusterReceiver, but can be breaking if the deprecated signalfx exporter
- Upgrade the Splunk OpenTelemetry Collector for Kubernetes dependencies #919,#909
- opentelemetry-operator upgraded to v0.37.0
- cert-manager upgraded to v1.12.4
- Enable native OTel logs collection (logsEngine=otel) in GKE Autopilot #809
- Configuration of persistent buffering for agent 861
- Add option to disable Openshift SecurityContextConstraint resource #843
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.83.0.
- Fix for secret name which now respects the same overrides as other resources in the chart #873
- Update the secret validation hook pod to use imagePullSecrets instead of possible non-existing serviceAccountName #888
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.82.0.
- Use "ContainerAdministrator" user for windows nodes by default #809
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.81.0.
- Set cluster_name for host logs too if renameFieldsSck is enabled #837
- Fix
smartagent
collectd based monitors on read-only filesystems #839
- Update PodDisruptionBudgets API version to allow both
policy/v1beta1
andpolicy/v1
#835 - Update clusterrole to allow collector to check for the
aws-auth
configmap in EKS clusters #840 - Add support to create default Instrumentation for operator based auto-instrumentation #836
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.80.0.
- Add
service.name
resource attribute to logs ifautodetect.istio
is enabled using transform processor. This change removes the limitation ofservice.name
attribute being available only with logsEngine=fluentd. #823 - Upgrade the Splunk OpenTelemetry Collector for Kubernetes dependencies #828
- Update the log level for metric scrape failures of the smartagent/kubernetes-proxy receiver from error to debug when distribution='' #832
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.79.1.
- Fix cri-o log format time layout #817
- Align the set of default resource attributes added by k8s attributes processor if the gateway is enabled #820
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.79.0.
- Adopt new control-plane node toleration #814
- Update the Kubernetes Proxy monitor for OpenShift clusters to start using secure ports 9101 or 29101 with authentication #810
This Splunk OpenTelemetry Collector for Kubernetes release adopts the Splunk OpenTelemetry Collector v0.78.1.
- Adopt Openshift distribution in Network Explorer #804
- Add
serviceAccountName
to secret validation hook pod #781 - Allow enabling profiling only for Observability #788
- Add storage to filelog receiver for extraFileLog #755
- Add Openshift support for Network-Explorer 804
- Avoid
runAsUser
in SecurityContext for Windows #797
- Option to use lightprometheus receiver through a feature gate for metrics collection from discovered Prometheus endpoints 757
- Add
logsCollection.containers.maxRecombineLogSize
config option with default 1Mb value which is applied tomax_log_size
option of the multiline recombine operators - Move
extraOperators
abovemultilineConfig
in _otel_agent.tpl, asextraOperators
gets skipped when we use bothmultilineConfig
ANDextraOperators
in values.yaml - Enable retry mechanism in filelog receiver to avoid dropping logs on backpressure from the downstream pipeline components #764
- Drop excessive istio attributes to avoid running into the dimensions limit when scraping istio metrics is enabled 765
- Fix k8s.cluster.name resource attribute for logs in GCP #771
- Update the Kubernetes scheduler monitor to stop using insecure port 10251 and start using secure port 10259 with authentication #711
- Upgrade splunk-otel-collector image to 0.75.0
- Sending request timeouts in k8s cluster receiver deployment on big k8s clusters #717
- Properly handle enableMetrics in Network Explorer Reducer template #724
- Add support for Operator based Java auto-instrumentation #701
- Add experimental support for deploying OpenTelemetry Operator as a subchart #691
- Improve documentation about providing tokens as Kubernetes secrets #707
- Expose
idle_conn_timeout
on the splunk HEC exporters #728 - Add
logsCollection.containers.maxRecombineLogSize
config option with default 1Mb value which is applied tomax_log_size
option of the default recombine operators #713
- Add functional test coverage for Network Explorer metrics #684
- Apply the same resources to init containers as allocated to the otel agent container #690
- Added examples for supported Kubernetes distributions and Kubernetes clusters with windows nodes (#663)
- Refactored the examples and rendered directories into one for better usability (#658)
- Docker metadata turned off by default (#655)
- Translation of
k8s.pod.labels.app
attribute to SCK format (#660)
- Support sending traces via Splunk HEC exporter (#629 - thanks to @mr-miles)
- Upgrade splunk-otel-collector image to 0.70.0, skipping 0.69.0 release (#653)
- Fix invalid OpenShift SecurityContextConstraints template (#652)
- Limit
clusterReceiver.eventsEnabled
deprecation warning to feature users (#648) - Fix noop validation for missing platform info (#649)
- Allow to overwrite default SecurityContextConstraints rules with values.yaml file (#643)
- Upgrade splunk-otel-collector image to 0.68.1 (#640)
- Default recombine operator for the docker container engine (#627)
- Added acl to journald log directory (#639)
- Upgrade splunk-otel-collector image to 0.67.0 (#612)
- Make sure the daemonset can start in GKE Autopiot (#608)
- Make containerd engine default in for fluentd logs and use always use it in GKE Autopiot (#609)
- Temporary disable compression in Splunk Observability logs exporter until 0.68.0 to workaround a compression bug (#610)
- Fixed network explorer image pull secrets
- Upgrade splunk-otel-collector image to 0.66.0 (#593)
- Upgrade splunk-otel-collector image to 0.64.0 (#589)
- Added Network Explorer components
- Make sure filelog receiver uses file_storage for checkpointing (#567)
- Upgrade splunk-otel-collector image to 0.62.0 (#573)
- Increase number of queue consumers in the gateway configuration (#554)
- Upgrade splunk-otel-collector image to 0.61.0 (#556)
- Upgrade splunk-otel-collector image to 0.59.1 (#536)
- [BREAKING CHANGE] Datatype of
filelog.force_flush_period
andfilelog.poll_interval
were changed back from map to string due to upstream changes. See upgrade guidelines
- [BREAKING CHANGE] Datatype of
- Make Openshift SecurityContextConstraints more restrictive (#513)
- Upgrade splunk-otel-collector image to 0.58.0 (#518)
- [BREAKING CHANGE] Datatype of
filelog.force_flush_period
andfilelog.poll_interval
were changed from string to map due to upstream changes. See upgrade guidelines
- [BREAKING CHANGE] Datatype of
- Do not send clusterReceiver metrics through gateway (#491)
- Upgrade splunk-otel-collector image to 0.57.0 (#504)
- Removed unnecessary change of group ownership in chmod initContainer (#486)
- Upgrade splunk-otel-collector image to 0.56.0 (#501)
- Upgrade splunk-otel-collector image to 0.55.0 (#485)
- The receiver.k8sclusterreceiver.reportCpuMetricsAsDouble feature gate has been removed (#487)
- If you are using this feature gate, then see the upgrade guidelines
- Remove high cardinality fields from k8s events: (#484)
- k8s.event.start_time
- k8s.event.name
- k8s.event.uid
- Make sure that logs are enabled to send k8s events (#481)
- Make sure that "sourcetype" field is always set on k8s events (#483)
- Fix failing cluster receiver with enabled profiling and disabled logs (#480)
- OTel Kubernetes receiver is now used for events collection instead of Signalfx events receiver (#478)
- Upgrade splunk-otel-collector image to 0.54.0 (#479)
- Fix recombining of oversized log records generated by CRI-O and containerd engines (#475)
- Fix bug where clusterReceiver splunk_hec exporter is enabled but configured not to send o11y logs (#471)
- A recombine operator for OTel logs collection to reconstruct multiline logs on docker engine (#467)
- Scrape /proc/self/mountinfo in agent pods to avoid incorrect stat attempts (#467)
- Upgrade splunk-otel-collector image to 0.53.1 (#468)
- Upgrade splunk-otel-collector image to 0.53.0 (#466)
- Add
splunkPlatform.retryOnFailure
andsplunkPlatform.sendingQueue
config options to values.yaml (#460)
- Upgrade splunk-otel-collector image to 0.52.2 (#463)
- Add troubleshooting documentation for incompatible Kubernetes and container runtime issues (#452)
- Fix native OTel logs collection where 0 length logs cause errors after the 0.29.0 opentelemetry-logs-library changes in 0.49.0 (#451)
- Upgrade splunk-otel-collector image to 0.51.0 (#453)
- Add gateway support for Host Logs (#437)
- Make sure that logs or profiling data is sent only when it's enabled (#444)
- Fix native OTel logs collection broken after the 0.29.0 opentelemetry-logs-library changes in 0.49.0 release (#448)
- Upgrade splunk-otel-collector image to 0.50.0 (#449)
-
Migrate filelog operators to follow opentelemetry-log-collection v0.29.0 changes (#436,#441)
- [BREAKING CHANGE] Several breaking changes were made that affect the filelog, syslog, tcplog, and journald receivers. Any use of the extraFileLogs config, logsCollection.containers.extraOperators config, and affected receivers in a custom manner should be reviewed. See upgrade guidelines
-
The receiver.k8sclusterreceiver.reportCpuMetricsAsDouble feature gate is now enabled by default (#487)
- [BREAKING CHANGE] The Splunk Otel Collector has a feature gate to enable a bug fix that makes the k8sclusterreceiver emit a few Kubernetes cpu metrics differently to properly adhere to OpenTelemetry specifications. See upgrade guidelines
-
Upgrade splunk-otel-collector image to 0.49.0 (#442)
- Upgrade splunk-otel-collector image to 0.48.0 (#434)
- Bug where the k8sclusterreceiver emits a few Kubernetes cpu metrics improperly (#419)
- [BREAKING CHANGE] The Splunk Otel Collector added a feature gate to enable a bug fix that makes the k8sclusterreceiver emit a few Kubernetes cpu metrics differently to properly adhere to OpenTelemetry specifications. See upgrade guidelines
- Upgrade splunk-otel-collector image to 0.47.1 (#422)
- Add support for otelcol feature gates to the agent, clusterReceiver, and gateway (#410)
- Upgrade splunk-otel-collector image to 0.46.0 (#413)
- Upgrade splunk-otel-collector image to 0.45.0 (#407)
- [BREAKING CHANGE] Use newer batch and autoscaling APIs in the Kubernetes cluster receiver (#433). The Kubernetes cluster receiver will not be able to collect all the metrics it previously did for Kubernetes clusters with versions below 1.21 or Openshift clusters with versions below 4.8. See upgrade guidelines
- Add environment processor to metrics pipeline when sending metrics to Splunk Platform (#399)
- Control plane metrics support: etcd (#384)
- Add missing splunk-otel-collector secret to gateway and cluster receiver deployment (#390)
- [BREAKING CHANGE] Set
profilingEnabled
to default false (#388)
- Added support to collect control plane component metrics; controller-manager, coredns, proxy, scheduler (#383)
- Explicitly set match_type parameter in filter processor (#385)
- Truncate eks/fargate cluster receiver StatefulSet names (#386)
- Support of profiling data for Splunk Observability (#376)
- [BREAKING CHANGE] OTel Collector Agent now overrides host and cloud attributes of logs, metrics and traces that are sent through it (#375). See upgrade guidelines
eks/fargate
distribution (#346)
- Upgrade splunk-otel-collector image to 0.43.0 (#370)
- Journald logs collection (#290)
- Automatic discovery and metrics collection from the Kubernetes API server control plane component (#355)
- Native OTel logs collection from the Windows worker nodes (#361)
- Option to disable helm hook for custom secret validation (#350)
- Upgrade splunk-otel-collector image to 0.42.0 (#367)
- Double expansion issue splunk-otel-collector (#357). See upgrade guideline
- Schema validation for
image.imagePullSecrets
configuration option (#356) - Schema validation for
logsCollection.containers.extraOperators
configuration option (#356)
- Temporary helper initContainer for OTel checkpointing log path move (#358)
- Google Kubernetes Engine Autopilot support (#338)
- Upgrade splunk-otel-collector image to 0.41.0 (#340)
- Upgrade splunk-otel-collector image to 0.40.0 (#334)
- Upgrade splunk-otel-collector image to 0.39.0 (#322)
- [BREAKING CHANGE] Logs collection is now disabled by default for Splunk Observability destination (#325)
This release completes the addition of content and documentation to easily allow users to send telemetry data including logs to both Splunk observability and Splunk platform. This will address use cases of current users of the Splunk Connect for Kubernetes.
Migration guidelines for Splunk Connect for Kubernetes users
Upgrade guidelines for existing Splunk OpenTelemetry Collector for Kubernetes users
- Field name compatibility for SCK (#258)
- Add initContainer for file operations for running as non root user (#263)
- Helm hook for custom secret validation (#294)
- Add include logs functionality based on pod annotations (#260)
- Support for tailing custom host files (#300)
- Extract
container.image.tag
attribute fromcontainer.image.name
(#285) - Upgrade splunk-otel-collector image to 0.38.1 (#284)
- Upgrade fluentd-hec image to 1.2.8 (#281)
- Change secret names according to the GDI specification (#295)
- Make
clusterName
configuration parameter generally required (#296) - Changed the default checkpoint path to
/var/addon/splunk/otel_pos
(#292) - Rename "provider" and "distro" parameters to "cloudProvider" and "distribution" (#297)
- Changed SplunkPlatform properties to match helm best practices. (#306)
- Rename parameter groups for Splunk OTel Collector components (#301):
otelAgent
->agent
otelCollector
->gateway
otelK8sClusterReceiver
->clusterReceiver
- Rename
stream
log attribute tolog.iostream
(#311) - Improve configuration for fetching attributes from annotations and labels of pods and namespaces (#273)
- Use
main
as default index and disable metrics by default for Splunk Platform (#305)
- Splunk Platform client certificates (#286)
logsCollection.containers.excludePaths
config parameter (#312)- Splunk Platform sourcetype precedence order (#276)
- Add initContainer for log checkpoint migration from Fluentd to Otel agent (#253)
- Add index routing for Splunk Enterprise/Cloud customers (#256)
- Fix metrics/logs disabling for Splunk Platform destination (#259)
- Fix kubernetes events in Observability IMM by adding
kubernetes_cluster
attribute (#261)
- Add recommended Kubernetes labels (#217)
- Add an option to skip RBAC resources creation (#231)
- Enable container metadata. This gives all collected logs new attributes:
container.image.name
andcontainer.image.tag
. Also the native OTel logs collection getscontainer.id
attribute that allows container level correlation in Splunk Observability Cloud closing a feature parity gap with fluentd (#238) - Add strict values.yaml schema validation (#227,#234,#239)
- BREAKING CHANGE: Reorder resource detectors, moving the
system
detector to the end of the list. Applying this change in an EC2 or Azure environment may change thehost.name
dimension and the resource ID dimension on some MTSes, possibly causing detectors to fire. - BREAKING CHANGE: Reduce scope of host mounted volumes on linux systems (#232)
- Change
run_id
log resource attribute tok8s.container.restart_count
(#226) - Use only
splunkPlatform.endpoint
andsplunkObservability.realm
parameters to identify which destination is enabled, remove default value forsplunkObservability.realm
(#230,#233) - Upgrade splunk-otel-collector image to 0.37.1 (#237,#249)
- Simplify configuration for switching to native OTel logs collection (#246)
- Fix setting of SPLUNK_MEMORY_TOTAL_MIB env var in otelAgent daemonset (#240)
- Enable OTLP HTTP ports (4318 and 55681) in otelAgent daemonset (#243)
- Exclude redundant
groupbyattrs/logs
processor from native logs collection pipeline (#219) - Fix deprecation messages for old
<telemetry>Enabled
parameters (#220)
- Fix backward compatibility for
splunkRealm
parameter (#218)
- Support k8s clusters with Windows nodes (#190)
- Change configuration interface to be able to send data to Splunk Enterprise/Cloud and to Splunk Observability (#209)
- Improve multiline logs configuration for native logs collection (#208)
- Add an option to provide additional custom RBAC rules (#206)
- Send k8s events additionally to Splunk HEC endpoint (#202)
- Remove PodSecurityPolicy installation option (#195)
- Add an option to collect k8s events with smartagent/kubernetes-events receiver (#187)
- Move k8s metadata enrichment from fluentd to otel-collector (#192)
- Upgrade splunk-otel-collector image to 0.31.0 (#183)
- Set more frequent checks for memory_limiter (#178)
- Make Fluentd init container image variable (#182)
- All missing attributes are added to prometheus metrics reported by gateway and k8s-cluster-receiver collector deployments (#170)
- Fix pod affinity setting (#181)
- Fix generation of service.name log attribute in istio environment (#176)
- Change internal metrics port from 8888 to 8889 (#172)
- Upgrade splunk-otel-collector image version to 0.29.0 (#174)
- BREAKING: Remove SAPM receiver from default config (#168)
- Upgrade splunk-otel-collector image to 0.28.1 (#166)
- Upgrade splunk-otel-collector image to 0.28.0 (#164)
- BREAKING CHANGE: Auto-detection of prometheus metrics is disabled by default (#163). See Upgrade guideline
- Fix container runtime detection when metrics pipeline disabled (#161)
- Add an option to add extra labels to pods (#158)
- Add an option to add extra annotations to deployments, daemonset, pods and service account (#158)
- Add an option to mount extra volumes to gateway-mode and k8s cluster receiver collectors (#157)
- Automatically detect container runtime using initContainers and apply relevant parsing config instead of asking user to specify criTimeFormat. This is an important change to enable smooth transition from deprecated docker to containerd runtime (#154)
- Add an option to mount extra volumes using
otelAgent.extraVolumes
andotelAgent.extraVolumeMounts
(#151)
- Add signalfx metrics receiver to the agent (#136)
- fluentd logs are now sent through the collector instead of being sent directly to the backend (#109)
- Logs are sent through the OpenTelemetry Agent on the local node by default.
otelAgent.enabled
value must be set totrue
when using logs (#127) otelAgent.ports
andotelCollector.ports
are selectively enabled depending on what telemetry types are enabled withmetricsEnabled
,tracesEnabled
, andlogsEnabled
- Removed setting
host.name
through theresource
processor as it is already set by theresourcedetection/system
processor - Upgraded to Splunk OpenTelemetry Collector 0.26.0
- Kubernetes cluster metrics now have a dimension
receiver:k8scluster
to ensure that MTS do not conflict with Kubernetes metrics sent by Smart Agent for the same cluster. (#134)
- Removed
ingestHost
,ingestPort
,ingestProtocol
, useingestUrl
instead (#123) - Removed
logsBackend
, configuresplunk_hec
exporter directly (#123) - Removed
splunk.com/index
annotation for logs (#123) - Removed
fluentd.config.indexFields
as all fields sent are indexed (#123) - Removed
fluentforward
receiver from gateway (#127) - Removed
service.ports
, sourced fromotelCollector.ports
instead (#140)
- Upgrade splunk-otel-collector docker image to 0.25.0 (#131)
- Pre-rendered manifests can be found in rendered directory
- Remove internal fluentd metrics sent as logs with monitor_agent. Prometheus metrics exposed on 0.0.0.0:24231 should be used instead (#122)
- Fix logs collection configuration for CRI-O / containerd runtimes (#120)
- Change the way to configure "concat" filter for container logs (#117)
- Disable fluentd metrics collection by default (#108)
- Change OTLP port from deprecated 55680 to default 4317 (#103)
- Open port for signalfx-forwarder on the agent (#106)
- Fix traces enrichment with k8s metadata (#102)
- Switch to stable Splunk OTel Collector image 0.24.3 (#100)
- Enable smartagent/signalfx-forwarder in the default agent trace pipeline (#98)
- Enable batch processor in the default metrics pipelines (#90)
- Ensure all metrics and traces are routed through the gateway deployment if it's enabled (#96)
- Add an option to set extra environment variables (#91)
- Fix resource attribute in the default traces pipeline (#88)
- Add metric_source:kubernetes for all k8s cluster metrics (#89)
- Fix host.name attribute in logs (#87)
- Add host.name attribute to logs (#86)
- Remove deprecated opencensus receiver (#85)