* Add retention policy for resolved alert history and diagnostics
- Introduce database retention for old resolved alert history and standalone diagnostics such as Explain Traces and orchestration execution traces.
- Add support for retention cleanup job with batching and interval configuration.
- Update models with new indexes to optimize retention cleanup queries.
- Extend settings to consolidate retention configuration in a unified `[retention]` section.
- Add tests for alert history cleanup behavior and edge cases.
- Update documentation with guidelines for configuring and understanding retention policies, including Russian translations.
* Publish Helm chart as OCI artifact and update release workflow
- Publish the IncidentRelay Helm chart to GHCR as an OCI artifact.
- Update GitHub Actions to include Helm chart validation, packaging, and publishing jobs.
- Improve documentation with installation instructions for OCI-based Helm chart usage.
- Add annotations to `Chart.yaml` for better metadata and compliance.
- Refactor existing workflows to address Helm chart publishing dependencies.
Closed #72
* Add configurable event processing trace levels through Event Orchestration.
- add set_trace_level orchestration action
- support full, compact and disabled trace modes
- keep full tracing as the default behavior
- buffer trace data until the effective trace level is known
- avoid creating explain trace rows when tracing is disabled
- omit detailed payload data in compact mode
- add trace_level and truncated metadata to explain traces
- bound buffered trace memory to prevent unbounded growth
- limit buffered trace size and step count per event
- discard detailed buffered data when limits are exceeded
- avoid building heavy orchestration explain payloads for compact/disabled traces
- show compact and truncated trace state in the UI
- expose the new action in Event Orchestration only
- reject set_trace_level in service-scoped orchestration
- add migrations, API/schema updates and regression tests
Closes #58
* Fix acknowledgement handling for correlated alert groups and prevent stale
notification work from paging responders after an incident has been ACKed.
- keep acknowledged incident_key groups sticky for same/lower-priority children
- reopen acknowledged groups only when a new child materially raises priority
- preserve manual and initial-only priority behavior
- reset escalation state when a higher-priority child legitimately reopens a group
- cancel pending notification, reminder and escalation deliveries on ACK
- make repeated acknowledgements clean up racing pending deliveries
- reload alert group state before processing reminder scheduler batches
- recheck persisted group status before notification delivery
- prevent stale notification queue workers from reverting ACKed groups to firing
- keep default reopen behavior for non-incident_key and multi-field grouping
- preserve orchestration group-key override semantics
- add regression coverage for ACK races and correlated incident behavior
* Add new relic integration
* Fix dark theme
* Refactor CSS and templates to adopt theme variables for improved dark mode support
* Add more frendly Orchestration simulator output
* Add a native Nagios inbound integration compatible with both Nagios Core
and Nagios XI.
- add Nagios route source and webhook endpoint
- normalize host and service notifications
- map PROBLEM notifications to firing alerts
- resolve alerts on RECOVERY, OK and UP states
- synchronize Nagios acknowledgements with IncidentRelay ACK lifecycle
- ignore downtime, flapping and custom notifications that should not create incidents
- add stable host/service deduplication keys
- map Nagios host and service states to IncidentRelay severities
- expose Nagios metadata as matcher-friendly labels
- add default grouping by nagios_host and nagios_service
- register the Nagios normalizer for production intake and simulation
- add Route UI, OpenAPI and localized documentation
- add a dependency-free Nagios notification sender example
- support secure route URL/token configuration through Nagios USER macros
- add regression coverage for trigger, recovery, acknowledgement and ignored lifecycle events
* Add config examples
