roxy-wi/IncidentRelay v2.2.0

on GitHub

Summary

IncidentRelay 2.2 adds alert shelving, native Azure Monitor support, expanded SSO profile mapping, configurable alert event history, additional localizations, and substantial lifecycle and reliability hardening.

Highlights

Alert Shelving

Alerts can now be temporarily shelved without changing their underlying firing or acknowledged state.

• configurable shelve duration and reason
• manual unshelve and automatic expiration
• pause notifications, reminders, updates, and escalation while shelved
• automatically resume delivery and escalation after unshelving
• safely clear shelving when alerts are resolved or merged
• Shelve / Unshelve actions in the web UI
• interactive shelving actions for Slack, Telegram, Mattermost, and browser push
• audit log and timeline events for shelving lifecycle changes
• filtering and visual indicators for shelved alerts

Service and business impact calculations continue to use the real alert state while an alert is shelved.

Azure Monitor Integration

Added native support for the Azure Monitor Common Alert Schema.

• Fired and Resolved lifecycle support
• Sev0–Sev4 severity mapping
• stable alert deduplication
• Azure metadata and custom properties exposed as routing labels
• route-token authentication
• HTTP Basic authentication for Azure Action Groups
• Azure Monitor route source and UI configuration
• OpenAPI coverage and integration documentation

Configurable Alert Event History

Alert event history can now be controlled globally or through Event Orchestration.

Available modes:

• full — keep the complete incoming alert lifecycle history
• initial — retain only the initial incoming lifecycle event
• disabled — suppress incoming lifecycle history

Operational timeline events such as acknowledgements, comments, escalations, maintenance changes, responders, correlation, and notifications remain available.

Event Orchestration can override the effective history mode with the new set_alert_event_history action.

Alert history also supports pagination and loading older events on demand.

SSO Profile Claim Mapping

SSO providers can now populate IncidentRelay messaging profile IDs directly from OIDC or SAML claims.

Supported profile fields:

• Slack user ID
• Telegram user ID
• Mattermost user ID

Custom and dotted claim names are supported. Existing manually configured IDs are preserved and are not overwritten during SSO login.

Soft-delete and Resource Lifecycle Hardening

Resource deletion and restoration behavior has been substantially hardened.

Highlights include:

• restore soft-deleted SSO users safely on successful login
• prevent deleted users from regaining old memberships, tokens, notification rules, push subscriptions, or runtime state
• clean up ownership, responders, stakeholders, rotations, and access assignments
• expanded group/team delete cascades
• revoke API and orchestration intake tokens when parent resources are deleted
• cancel pending orchestration and notification work
• resolve active alert groups belonging to deleted teams
• reset heartbeat runtime state during deletion
• prevent scheduler and notification workers from using deleted or inactive resources
• safely recreate resources with previously used natural keys while preserving historical records

Localization

Added two new complete UI localizations:

• Spanish
• Simplified Chinese

Chinese localization includes profile preference support, browser locale negotiation, PWA messages, calendar formatting, and dynamically generated Services UI text.

Helm and Deployment

• added support for custom CA certificates
• documented private-network outbound HTTP allowlisting
• fixed Helm numeric configuration rendering that could produce scientific notation for integer values
• expanded Helm regression coverage
• fixed redundant static CSS paths for deployments behind HTTPS/reverse proxies

API, Documentation, and Reliability

• expanded OpenAPI coverage for new and updated endpoints
• improved browser push and notification action handling
• additional Slack, Telegram, and Mattermost lifecycle hardening
• additional Event Orchestration validation and runtime safeguards
• expanded edge-case and regression test coverage
• updated English and Russian documentation

Database Migrations

This release contains database migrations for:

• SSO profile claim mappings
• alert shelving

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

Upgrade Notes

When upgrading from 2.1:

1. Back up the IncidentRelay database.
2. Review the licensing changes.
3. Review new configuration options, especially alert event history and outbound/custom CA settings.
4. Run the standard database migrations.
5. Upgrade the application and any scheduler / notification worker processes together.

Existing installations retain normal alert processing behavior unless the new functionality is explicitly configured.