roxy-wi/IncidentRelay v2.1

on GitHub

Summary

Version 2.1 significantly expands IncidentRelay's event processing, orchestration, integrations, incident lifecycle, operational visibility, deployment options, and frontend UX.

Highlights

Event Orchestration

• expanded Event Orchestration capabilities and actions
• added deterministic orchestration simulation
• improved Simulator UX with:
  • human-readable summary
  • matched/skipped rule evaluation
  • input → output changes
  • active version → draft comparison
  • raw JSON view
• added execution history and detailed execution trace modal
• load full execution traces on demand instead of loading traces for the entire execution list
• added configurable Explain Trace levels:
  • full
  • compact
  • disabled
• added bounded Explain Trace buffering to prevent unbounded per-event memory growth

Alert and incident lifecycle

• improved alert grouping, correlation and priority handling
• preserve acknowledgement for correlated incident_key alerts when effective priority does not increase
• correctly reopen acknowledged incidents on material priority escalation
• reset escalation state when an acknowledged incident legitimately reopens
• cancel pending notifications, reminders and escalations after ACK
• protect notification/reminder workers from stale group state races
• improved maintenance and silence lifecycle handling
• expanded alert processing Explain Trace and timeline visibility

Integrations

Added and improved inbound integrations, including:

• New Relic
• Nagios Core / Nagios XI
• Datadog
• LibreNMS
• Zabbix
• Grafana
• Alertmanager
• Sentry
• AWS SNS / CloudWatch
• Uptime Kuma
• RMON
• generic webhook / PagerDuty Events API v2 compatibility

Nagios support includes host/service lifecycle normalization, recovery handling, acknowledgement synchronization and a dependency-free notification sender example.

Database retention

Added configurable database retention for historical event-processing data.

Retention can be configured independently for:

• resolved alert history
• Explain Traces
• Event Orchestration executions

Cleanup is batched and protected by a distributed database lock.

0 disables retention cleanup and keeps records indefinitely.

Helm and deployment

• added OCI publication support for the Helm chart
• chart can be installed directly from GHCR
• aligned chart/application release version checks
• improved release validation and documentation

Example:

helm install incidentrelay \
  oci://ghcr.io/roxy-wi/incidentrelay-charts/incidentrelay \
  --version 2.1.0

Services and operational visibility

• expanded service dependency and business-impact functionality
• improved Impact v2
• improved service context and dependency visualization
• expanded heartbeat monitoring and instance handling
• improved alert/service timelines and diagnostic information

Frontend and UX

• reworked the application-wide dark theme around centralized semantic theme tokens
• removed duplicated page-specific dark theme rules
• fixed light surfaces and low-contrast components across the application
• improved dark mode for modals, tables, forms, services, impact views, heartbeats, SSO and alert details
• made dependency graphs and charts theme-aware
• added URL-backed state for filters, search, sorting and list views
• preserve list state across refresh, browser navigation and shared URLs
• moved Route details into a modal
• redesigned Route details into a responsive multi-column layout
• improved modal and table consistency throughout the UI

SSO and access management

• expanded SSO mapping capabilities
• improved group/team mapping behavior
• fixed editing of OIDC mapping rules
• improved SSO mapping UI and role handling

API, documentation and localization

• expanded OpenAPI coverage for new and updated APIs
• updated English and Russian documentation
• expanded UI localization coverage

Database migrations

This release contains database migrations.

Back up the database before upgrading and run the normal IncidentRelay migration procedure as part of the deployment.

Upgrade notes

Review the new configuration options, particularly:

[retention]
alert_days = ...
explain_trace_days = ...
orchestration_execution_days = ...
cleanup_interval_seconds = ...
batch_size = ...

Existing installations keep the current full Explain Trace behavior unless explicitly configured otherwise.