Feature Changelog

IncidentRelay Changelog.

Release changes

v2.5.1

  • Preview and schedule one-way imports of teams, users and on-call schedules from external APIs with configurable field mappings and update rules.
  • Explore alert activity with previous-period comparisons, a weekday/hour heatmap and the largest increases in alert noise. Open matching alert groups directly from each result.
  • Send native GitLab Deployment events to Change Intelligence and map projects and environments to Services. Change intake and automatic correlation remain opt-in; deployments are not monitoring alerts.
  • Manage Change integrations and system settings with appropriately scoped personal API tokens, subject to administrator and Group/Team restrictions.
  • Keep service and involvement filters synchronized between Dashboard and Alerts, including after a page reload.
  • Reduce repeated database work during alert intake and improve PostgreSQL connection pooling. Independent alert groups can deliver notifications concurrently on PostgreSQL; SQLite delivery remains sequential.
  • Strengthen API token scope and Group/Team checks, route-level deduplication boundaries and protection for outgoing SAML metadata and browser-push requests.
  • Handle voice ACK/resolve callbacks during delivery more reliably. Startup checks detect encryption-key changes, and Docker readiness checks verify database connectivity and migration state.

Start with the directory synchronization guide and GitLab deployment setup.

Release changes

v2.5.0

  • Coordinate an outage on one Incident: name the commander, request responders, review stakeholders and record operational comments, root cause and the resolution summary.
  • The Incident workspace brings together affected services, accessible dependencies, runbooks and linked technical activity. AlertGroups keep technical comments, paging, acknowledgement and escalation.
  • Review deployments and other recorded changes alongside an AlertGroup or Incident. The Changes Inbox, suggested links and manual review help investigate timing; a link is not proof of root cause.
  • Global administrators can independently enable Change Event intake or pause background processing without restarting IR. Twelve advanced limits are editable under System settings; edits are audited. Intake is off by default.
  • Receive Google Cloud Monitoring opening and closure notifications directly via a webhook. Enable closure notifications on the Google alerting policy; no Pub/Sub or Cloud Function is required.
  • Import the Grafana operations dashboard to inspect IR request latency, worker heartbeats and notification queues using your existing Prometheus scrape.
  • Breaking API change: responder and stakeholder endpoints now require a real Incident ID. The old AlertGroup endpoints do not forward requests.

Before upgrading, back up the 2.4 database and reconcile legacy participants. Records without exactly one eligible linked Incident are skipped. Read the upgrade and rollback guide first.

Release changes

v2.4.0

  • OpenSearch Alerting and Elastic Kibana Alerting can now send individual alerts and explicit recoveries to native webhook routes.
  • An optional reopen window lets a recurring fault reuse its resolved AlertGroup with a new child alert. A linked Incident in monitoring or resolved can return to investigating; closed Incidents stay closed.
  • Delayed recovery notifications retry failed recipients independently. A quick reopen cancels outstanding recovery deliveries for the previous occurrence.
  • Email channels can notify the assignee, use a fallback mailbox when the assignee has no address, or always send to a fixed mailbox.
Release changes

v2.3.1

  • Alert analytics adds noise rankings, acknowledgement rates, MTTA and MTTR, lifecycle trends and the oldest unresolved alert groups.
  • Prometheus metrics expose request latency, alert intake, notification queues and worker health, with optional bearer-token authentication and Helm ServiceMonitor support.
  • Configuration supports environment-variable overrides, mounted secret files and Helm references to existing Secrets or ConfigMaps.
  • Orchestration execution traces show matched rules, event changes and candidate decisions versus actual lifecycle results.
  • Alert details show the current priority policy and notification-routing source, while incident linking can search alert groups by title or ID.
  • Docker images now support linux/arm64 alongside linux/amd64.
Release changes

v2.3.0

  • Incidents now have their own workspace, lifecycle, priority and operational owner, with alert-group linking and an activity timeline.
  • Breaking API change: technical alert-group operations move to /api/alert-groups; /api/incidents now manages operational Incidents only.
  • Cloud.ru Cloud Eye integration receives signed SMN alarm and recovery notifications with automatic subscription confirmation.
  • Feishu / Lark channels deliver text notifications through custom bot webhooks with optional request signing.
  • The Overview dashboard adds alert trends, service health, team activity, current on-call responders and recent activity.
  • Notification policies gain filters for priority, severity, source and service attributes, while preserving disabled channel references.
  • Mobile alert details and browser-push navigation are improved, and Grafana webhooks accept optional fields set to null.
Release changes

v2.2.0

  • Alert shelving can temporarily pause notifications, reminders, updates and escalation for one active alert group without hiding its technical state or service impact.
  • Responders can shelve and unshelve alerts from the web interface, browser push, Slack, Telegram and Mattermost, with automatic expiry, resumed delivery and complete timeline and audit records.
  • A native Azure Monitor integration accepts Common Alert Schema webhooks, maps Fired and Resolved events, normalizes Sev0-Sev4 priorities and exposes Azure metadata for routing.
  • Incoming child-alert event history can use full, initial-only or disabled storage globally, while Event Orchestration can override the mode for matching events without removing operational timeline entries.
  • OIDC and SAML providers can map identity claims to Slack, Telegram and Mattermost user IDs, making interactive notification actions easier to provision through SSO.
  • Resource deletion and recreation are safer: dependent routing and scheduler work is disabled consistently while historical records remain available.
  • The interface adds complete Spanish and Simplified Chinese localizations.
  • Helm deployments gain custom CA support and safer private-network outbound HTTP configuration, with fixes for numeric values and reverse-proxy static paths.
Release changes

v2.1

  • Configurable retention policies can automatically remove old resolved alert history, Explain Traces and Event Orchestration execution traces in bounded scheduler batches.
  • Event Orchestration can set full, compact or disabled processing traces per matching event, with bounded buffering and visible compact or truncated state.
  • Acknowledging a correlated incident now cancels stale notification, reminder and escalation work; same or lower-priority children stay acknowledged while higher-priority events can reopen the group.
  • A native New Relic integration receives Alerts Workflow issue notifications and routes normalized lifecycle events through standard IncidentRelay intake tokens.
  • A native Nagios Core and Nagios XI integration maps host and service problems, recoveries and acknowledgements while ignoring downtime, flapping and custom notifications.
  • The Helm chart is now published to GHCR as an OCI artifact, with hardened defaults and improved SQLite-safe deployment checks.
  • Dark-theme styling is more consistent across core views, and the Orchestration simulator presents results in a clearer, friendlier format.
Release changes

v2.0

  • Event Orchestration adds global and service-scoped, versioned rules that can normalize, enrich, route, group and deduplicate incoming events before the alert lifecycle finishes processing them.
  • Rules can select teams, routes, services and policies, change priority and severity, suppress notifications, delay or drop events, and queue secure asynchronous webhooks.
  • A visual builder and JSON editor are backed by validation, simulation, replay, shadow execution, traces, metrics and immutable publish-and-rollback workflows for safer rollouts.
  • A native Uptime Kuma integration turns monitor state changes into routed alerts and automatically resolves them when monitors recover.
  • Silences and maintenance windows can now apply retroactively to unresolved alerts; affected alerts can be reactivated when suppression ends, while escalation and notification state is kept consistent.
  • The interface adds a dark theme, saved UI preferences, an audit-log view and complete French localization.
  • Personal API tokens gain granular scopes, while login throttling, safer outbound HTTP requests and stricter validation strengthen the security boundary.
  • A bundled Helm chart adds a supported path for deploying the web app and background workers to Kubernetes.
Release changes

v1.2.1

  • PostgreSQL migrations now resolve tables, columns and indexes against the schema selected by the active search path, fixing upgrades that use a non-public schema.
  • Shared schema-aware introspection helpers are used across legacy migrations and schema checks, avoiding unsafe fallback to the public schema.
  • Migration upgrades and rollbacks now keep schema changes and migration records in the same database transaction, reducing the risk of partially applied migrations.
  • New migration tests cover custom PostgreSQL schemas, missing relations, index lookup, database proxies and non-PostgreSQL behavior.
Release changes

v1.2

  • IncidentRelay now includes UI localization support with English, Russian and German catalogs across alert, service, route, rotation, profile and administration workflows.
  • Grafana OnCall and PagerDuty migration tools help teams move existing on-call data into IncidentRelay.
  • PagerDuty Events API v2-compatible and Datadog webhook integrations expand the supported alert sources.
  • Slack Socket Mode support enables Slack app interactions without exposing a public inbound Slack endpoint.
  • Channel configuration secrets are masked more consistently in API and UI responses.
Release changes

v1.1

  • Official IncidentRelay container images are now published to GitHub Container Registry, and Docker Compose uses the versioned release image by default.
  • The Helm chart now ships as version 1.1.0 and uses the official IncidentRelay image for Kubernetes deployments.
  • Audit records, validation errors and readiness responses now redact secrets and avoid exposing sensitive input or database details.
  • Repeated acknowledge and resolve actions are idempotent, preventing duplicate timeline events, notifications and impact updates.
  • Service links and runbooks now validate external URLs before rendering actionable links.
Release changes

v1.0.33-beta

  • Heartbeats now support multi-instance tracking for jobs or agents running across many producers.
  • Auto-discovery can learn new heartbeat instances from successful pings, while static mode can enforce an expected producer list.
  • Each overdue heartbeat instance can create its own routed alert with instance labels and independent auto-resolve behavior.
  • Heartbeat instance state is exposed in the UI and API, including listing and disabling obsolete producers.
Release changes

v1.0.32-beta

  • Heartbeats add dead-man-switch monitoring for watchdogs, backups, ETL jobs, log shippers and other systems where silence is the failure mode.
  • Interval and scheduled-completion modes can create normal routed IncidentRelay alerts when expected pings stop arriving.
  • Heartbeat alerts use the existing route, escalation, notification, maintenance and service-impact pipeline, with optional auto-resolve when pings return.
  • Business Services and Service Catalog impact scoring now handle status, priority, severity and dependency weights more granularly.
Release changes

v1.0.30-beta

  • Business Services model customer-facing capabilities composed of one or more technical services.
  • Calculated status uses effective component health, dependency impact, component criticality and configurable impact weights.
  • Manual status overrides can describe customer impact and optionally expire automatically.
  • Business Service details include components, active incident impact and status history, with full UI and API management.
Release changes

v1.0.29-beta

  • Dependency-aware correlation links related active alert groups across service dependencies.
  • Possible root causes, symptoms and downstream impact include a confidence score, dependency depth and explanation.
  • Correlation appears in alert lists, details, timelines and notification context while remaining advisory: alerts are not merged, suppressed or resolved automatically.
  • Per-dependency settings and propagation delays control correlation timing and noise.
Release changes

v1.0.28-beta

  • Users can opt into Mattermost direct messages when their on-call shifts begin.
  • The Mattermost bot API now supports personal shift-start notifications.
  • Delivery deduplication prevents repeated notifications for the same shift.
  • SSO team-role mappings improve role assignment during login and user provisioning.
Release changes

v1.0.27-beta

  • Service Impact History stores scheduled or manual point-in-time snapshots of affected services, alert impact, dependency impact, root causes and blast radius.
  • Historical charts show affected-service trends and impact reasons, while rankings identify services that are repeatedly affected.
  • Service reliability analytics combine current SLI/SLO health with historical impact data for operational reviews and incident investigation.
  • Service context management now provides clearer switching and editing workflows for links, runbooks and dependencies.
Release changes

v1.0.26-beta

  • Services can define SLIs and SLOs for acknowledgement latency, resolution latency, incident-based availability and incident counts.
  • Rolling evaluations classify objectives as met, at risk, breached or without data and calculate availability error budgets.
  • SLOs support severity and priority scopes, configurable windows, open incidents and optional maintenance-window exclusion.
  • Service details, aggregate analytics and the service timeline expose current measurements and configuration changes.
Release changes

v1.0.24-beta

  • Service Notification Policies select shared delivery channels through ordered rules based on alert matchers and event types, with route-only, service-policy and combined delivery modes.
  • Priority Policies automatically assign P1-P5 using ordered matcher rules, team defaults or service overrides, configurable update modes and source or fallback priorities.
  • Reusable Matcher Presets and a shared editor bring consistent JSON validation, formatting and suggestions from recent alerts to routes, service rules, silences, policies and service runbooks.
  • Responders can reset a manually assigned incident priority to automatic management, while Alert Explain Trace shows how priority was selected and applied.
  • Email, Slack, Mattermost, Telegram and browser push notifications can link directly to the original source event, with improved Sentry event and project detection.
Release changes

v1.0.22-beta

  • On-call Health adds readiness indicators to Rotations and Teams, highlighting healthy, warning and critical configurations before an incident occurs.
  • Detailed diagnostics detect empty or disabled rotation layers, inactive members, schedule gaps, unresolved assignment targets and notification delivery problems.
  • Health details include issue counts, the checked time window and actionable recommendations for repairing each affected team or rotation.
  • New summary and detail API endpoints expose stable machine-readable issue codes for automation, monitoring and external diagnostics.
Release changes

v1.0.21-beta

  • Alert Explain Trace records how incoming alerts are processed, including route matching, grouping, maintenance and silence decisions, notification scheduling and stopped processing paths.
  • Integration ingest responses now include a trace_id, making routing failures, suppressed incidents and orphan resolved alerts easier to investigate from the UI or API.
  • The Alerts page includes an Explain view and deep links for opening traces directly from trace_id query parameters.
  • The Alerts API now validates list queries through a dedicated schema, improving multi-value filters, pagination, sorting and assigned-to-me query behavior.
Release changes

v1.0.20-beta

  • LibreNMS API Transport alerts can now be received through a dedicated librenms integration endpoint with normalized status, severity, labels, deduplication and recovery handling.
  • Users can set a profile timezone preference, improving alert and calendar time rendering for distributed teams.
  • The Notification Center exposes pending responder requests and user-facing incident actions through a dedicated API.
  • Services API documentation now covers inventory, match rules, links, runbooks, dependencies, impact and analytics endpoints in one dedicated reference.
  • The Alerts page can preserve filter state and includes an assigned-to-me filter for faster personal triage.
Release changes

v1.0.19-beta

  • A Helm chart is now available for Kubernetes deployments, including web, scheduler and Telegram worker components, rendered config secrets, probes, ingress, service account and persistence values.
  • Services can define default stakeholders that are copied into new incidents, so owners and observers can receive created, priority, status and resolved notifications.
  • Incident responder requests now feed a notification center where targeted users can see pending requests and accept or decline them from the incident workflow.
  • Incident and service APIs now cover responders, stakeholders, priorities, services and service owner/default stakeholder management.
Release changes

v1.0.17-beta

  • Incident priorities now use a consistent P1 Critical through P5 Informational scale across alert details and operational views.
  • The Alerts page supports priority filtering and sorting, while the dashboard shows priority badges and a priority distribution summary.
  • Incoming alert severity can automatically assign or raise incident priority without overwriting a manually selected priority or lowering an existing automatic priority.
  • Email, Mattermost, Telegram and browser push notifications now include incident priority, and manual priority changes can notify subscribed stakeholders by email.
Release changes

v1.0.16-beta

  • Read-only CalDAV sync exposes accessible team on-call schedules to Apple Calendar, Thunderbird, DAVx5 and other compatible clients.
  • Tokenized ICS subscription feeds let teams add on-call schedules to Outlook, Google Calendar and other calendar clients, with controls to regenerate or revoke feed URLs.
  • A new signed Sentry integration receives issue alerts, metric alerts and issue lifecycle events, with route-level secrets, deduplication and automatic resolve handling.
  • Route setup now provides source-specific webhook URLs, token details and ready-to-copy curl examples.
  • Health and readiness endpoints support Kubernetes probes and load balancer checks.
Release changes

1.0.15-beta

  • Service Impact v2 shows current effective service status, root causes, dependency paths and downstream blast radius.
  • Service Analytics v2 combines grouped incidents, raw alert volume, noise and deduplication metrics, maintenance counters and response-time data.
  • The Services experience now includes richer impact and analytics views with charts and clearer operational summaries.
  • User selection controls across teams, groups, rotations and escalation workflows have clearer labels and dynamic updates.
Release changes

1.0.14-beta

  • Maintenance Windows can target groups, teams, services or routes and suppress notifications, suppress new incidents, create maintenance incidents or pause escalations.
  • Maintenance schedules support timezones, effective statuses and recurring RFC5545 rules.
  • Responders can create, edit and delete investigation comments directly on alert groups, with actions reflected in the alert timeline.
  • Alert details now include incident management controls for priority, responder requests and stakeholder notification preferences.
  • Group managers can manage team users within their access scope.
Release changes

v1.0.12-beta

  • Related monitoring signals can now be collected into alert groups using configurable route group_by fields.
  • Operators can acknowledge, resolve and manually merge alert groups while retaining their individual child alerts.
  • Alert group notifications support delayed batching and update intervals, reducing noise from short-lived or rapidly expanding incidents.
  • Rotation layer members now have start and end periods, preserving historical schedules and supporting future membership changes.
  • Alert group queries support multi-value filters for more flexible incident views and API requests.
  • Administrators can create or update IncidentRelay admins from a Slack usergroup with the new sync-slack-admins command.
Release changes

v1.0.11-beta

  • Browser/PWA push notifications are now available for assigned users from their profile devices.
  • Browser push notifications can include short-lived ACK and Resolve actions.
  • Profile notification rules let users define follow-up delivery through browser push, email or voice call.
  • Notification delivery now handles skipped delivery checks and browser push behavior more reliably.
  • Alerts and admin-style tables now use reusable filtering and pagination helpers for larger datasets.
Release changes

v1.0.10-beta

  • On-call status is now visible in the topbar and profile experience, so users can see current and upcoming shifts faster.
  • Profile settings now include on-call status and notification preferences for shift scheduling email updates.
  • Shift email notifications can warn users about upcoming on-call periods while respecting preferences and deduplication rules.
  • Rotation lookups now account for layers, overrides and per-user participation when calculating user-specific on-call windows.
  • Calendar and rotation views were refined around time-bounded schedule lookups and clearer on-call context.
Release changes

v1.0.9-beta

  • Service management now models affected systems with owners, criticality, environments, metadata and default response paths.
  • Service match rules can map incoming alerts to services from route defaults, labels, annotations and payload fields.
  • Service dependencies and impact metadata now show how upstream services affect related systems and open alerts.
  • Integration-provided runbooks can be attached from alert payloads and shown before service-level runbooks.
  • The Services UI and API expose service inventory, match rules, dependencies and impact context for operations workflows.
Release changes

v1.0.8-beta

  • Escalation policies now support multi-step alert escalation chains per team.
  • Routes can use escalation policy mode instead of assigning directly through a single rotation.
  • Policy rules can target rotations or specific users with ordered delays and repeat counts.
  • The Escalation Policies UI and API let teams create, edit and inspect policies and policy rules.
  • Rotation calendar and override workflows were refined around the new escalation model.
Release changes

v1.0.7-beta

  • SSO authentication now supports OIDC and SAML 2.0 providers.
  • Administrators can configure SSO providers, login policies, allowed domains and external group mappings from the new SSO page.
  • The login page can show enabled SSO provider buttons and redirect users through the selected identity provider.
  • SSO login can auto-create or link users and synchronize external groups to IncidentRelay group roles.
  • SAML metadata, signing options and encrypted SSO secrets are now documented and exposed through the SSO flow.
  • The application UI received a Material-style refresh with cleaner navigation and improved profile/routes interactions.
Release changes

v1.0.6-beta

  • Rotations now support layered on-call schedules with priorities, active time restrictions, per-layer members and timezone-aware handoff rules.
  • The calendar now renders the final schedule from layers and overrides, so teams can inspect the effective on-call user for each time segment.
  • Calendar filtering now supports selecting a specific rotation when a team has multiple schedules.
  • Rotation APIs now expose layer, layer member and layer restriction endpoints for automation and integrations.
Release changes

v1.0.5-beta

  • Rotation reminder intervals can now be set to 0 to disable repeated reminders for alerts assigned through that rotation.
  • Reminder-based escalation now respects disabled reminders instead of falling back to a global reminder interval.
  • The Alerts page severity filter now includes warning and info severities for more precise triage.
  • Personal API token lifetimes are now capped at 365 days while still allowing non-expiring tokens with 0 days.
Release changes

v1.0.4-beta

  • Notification delivery now uses dedicated notifiers for email, Mattermost, Telegram and voice call channels.
  • Mattermost Bot API notifications can create and update alert posts with status-aware messages and action buttons.
  • Voice call notifications now use the configured voice provider with spoken alert text, callback support and DTMF action mapping.
  • Channels API documentation now includes supported channel types, the default email template endpoint and clearer channel schemas.
  • The Silences page can filter from summary cards and optionally include expired history older than 30 days.
  • Docker and systemd documentation now use the unified service name incidentrelay instead of incidentrelay-web.
Release changes

v1.0.3-beta

  • Email notifications can now use customizable HTML templates for richer and more consistent alert messages.
  • Email channel settings include template controls and supported placeholders directly in the UI.
  • Email delivery is more reliable when teams use custom templates or global SMTP settings.
  • User administration now uses clearer role names and more detailed role management workflows.
Release changes

v1.0.2-beta

  • Team and group lists now focus on active records, reducing clutter in day-to-day administration.
  • Validation is more consistent across API forms, so invalid values are caught earlier and reported more predictably.
  • Phone number fields are easier to fill correctly thanks to stricter validation, normalization and immediate frontend feedback.
  • Login and form status messages now give clearer feedback during common setup and administration flows.
Initial beta baseline

v1.0.0-beta

  • Alerts now have shareable detail pages, making it easier to send a direct incident link to teammates.
  • Escalation progress is visible directly on alerts, helping responders understand how far an incident has escalated.
  • Alert state is easier to track through clearer resolved and escalation-level data in the API.
  • Telegram notifications can point back to the related alert and show the right action buttons for the current state.
  • Telegram support includes rich messages, interactive actions, inline buttons and polling mode.
  • Notification channels can use severity filters with normalized severity handling for more predictable routing.
  • On-call overrides are easier to manage with a reusable modal, faster cached state and calendar-aware behavior.
  • Administration, profile, calendar, rotation and alert screens were redesigned for clearer daily operations.