Observe Agent updates¶
The Observe Agent is released periodically and not on the same timeline as the Observe platform.
Version 2.10.0¶
Features
feat: add RED metrics config options for dimensionality and filtering
feat: add auth logging headers to heartbeat exporter
feat: add more facets + fields to heartbeat event
feat: add omit_base_components config option to use minimal bundled config
feat: add optional cumulative to delta processor for forwarded OTLP metrics
feat: add resource/agent_instance processor to all logging and internal metrics when fleet is enabled
feat: implement config heartbeat event
feat: set env var OBSERVE_AGENT_ENVIRONMENT to the OS/bundled config name
feat: split out logs for observe agent in windows and add agent instance
feat: update golang to 1.24.9
feat: upgrade golang to 1.24.8
feat: upgrade to OTel version 0.138.0
Bug fixes
fix: remove agent_instance from other unrelated pipelines
fix: use configured port for self monitoring metrics, clean up bundled config templates
Others
chore: switch template functions to Sprig to match Helm syntax
chore: update goreleaser version
Version 2.9.1¶
Features
feat: Add heartbeat receiver configs
feat: Add restart on install to linux postinstall/postremove scripts
feat: Heartbeat Receiver + ENV Var logic
feat: add integration test that covers upgrade path
feat: add jmx receiver
feat: bump to otel col version 0.135.0
Bug fixes
fix: Continue setting env vars even when agentresource creation fails
fix: Revert “fix: restart agent on restart on debian”
fix: have agent restart on upgrade on debian
fix: restart service on debian and RHEL
fix: update postinstall script syntax
Others
Move timeout from 25 minutes to 35 minutes
chore: bump alpine version, address vulnerabilities
chore: ignore vendor files and test data in orca scans, clean up warnings
chore: update windows test AMI IDs
Version 2.9.0¶
Features
feat: bump to otel col version 0.135.0
Version 2.8.1¶
Bug fixes
fix: pin prometheus version to fix observe-agent metric output format and status check
Version 2.8.0¶
Features
feat: add RED_metrics flag to init-config
feat: add config option for otlp receiver endpoints
feat: add metricstransformprocessor
feat: upgrade otel to v0.131.0
Bug Fixes
fix: make mac install script executable
fix: remove setting of stable feature gates, add feature gate error logging
fix: update github.com/docker/docker package to fix vulnerability
fix: updated gh workflows to go 1.24.6
Others
fix OTel version in the README
update golang version to 1.24.6
Version 2.7.2¶
Bug fixes
fix: always default span and RED metrics service resource attributes to empty string
fix: improve naming of RED metric attributes
fix: use proper attribute name in ottl check for observe.status_code
Version 2.7.1¶
Bug fixes
fix: use env as default config resolve schema
Version 2.7.0¶
Features
feat: add an option to generate RED metrics from spans.
The traces/forward pipeline now defaults to dropping spans with duration longer than one hour to match Observe Trace Explorer behavior. This can be configured via the
forwarding.traces.max_span_durationconfig option.
Version 2.6.0¶
What’s new in version 2.6.0:
JSON schema for our
observe-agent.yamlconfig file. This makes editing the agent config much easier and safer!New
config validatecommand which we can now use to validate our helm chart output.Added the
groupbyattrsprocessor.All bundled config is now part of the build, so no need for all of the extra template files on disk.