This version
This is the documentation for the released OpsKnight 2.2.0 product.
v2.2.0 is the current release and is listed in releasedVersions in
docs/versions.json, so the release-gated documentation sync publishes it.
Earlier releases keep their own documentation sets (for example docs/v2.1.0),
published under their own version.
The baseline documentation is grounded in product evidence. Historical documentation can help locate a topic, but it is never authoritative.
New in 2.2.0
Upgrading from 2.1.x? Read Upgrade to 2.2.0 first: it covers the migrations, the HTTPS default for outgoing webhooks, older integration keys, the 60-minute default snooze, known limitations, and rollback.
-
AI assistants (MCP), off by default: overview, self-hosted setup, the stdio bridge, what to ask, build an agent, security and data, troubleshooting, permissions, the tool reference, and for operators the architecture, monitoring, and release status.
-
Guided single sign-on: single sign-on guides, including role and team rules and linking existing accounts, plus sign-in troubleshooting.
-
Maintenance windows: hold alerts during planned work, one-off or recurring.
-
Dark mode: choose System, Light, or Dark under profile and preferences.
-
Integrations: Slack
/incidentnow needs thecommandsscope; copied integration URLs take their hostname from Settings → System → App URL first. -
Helm: the pre-upgrade runtime gate no longer stops every Deployment for values-only upgrades.
-
Analytics rebuilt: Analytics uses one exact metrics module, with sections, ranges up to 3 years, comparison, time zones, metric definitions and the CSV export (use Analytics).
An organisation-wide workflow builder and alert correlation (AIOps) are not part of 2.2.0. Workflow automation in this release is per-service automation.
Start here
- Evaluate with Docker Compose
- Create and resolve a first incident
- Plan a production installation
- Migrate from a 1.x deployment
Respond and operate
-
Understand service automation, configure and roll it out, and review its exact contracts.
-
Understand Runbook safety and follow the Runbooks guides to author, bind, and execute workflows.
-
Install outbound Agents, manage pools and scoped secrets, and diagnose execution issues.
-
Create, acknowledge, assign, and resolve incidents.
-
Build an on-call schedule and manage temporary overrides.
-
Publish a status update and create a ChatOps war room.
Connect systems
-
Browse the integration catalog for monitoring, cloud, uptime, source-control, communication, issue-tracking, and generic webhook providers.
-
Configure Slack, Microsoft Teams, voice paging, or Jira.
-
Operate service automation and diagnose automation problems.
Run in production
- Choose a deployment topology, then follow the complete Compose, Kubernetes, or multi-node path.
- Plan scaling, hardening, backup and restore, and upgrades, starting with Upgrade to 2.2.0.
Exact contracts
- API inventory
- Configuration
- Permissions
- Notification delivery
- Webhooks
- Health, metrics, and runtime limits
Evidence order
When sources conflict, use this order:
- Current executable behavior
- Current tests
- Current source code and configuration
- Historical documentation
Documentation must not claim behavior that cannot be verified against one of the first three sources.
Documentation sections
- Start — install OpsKnight and complete a first incident.
- Concepts — understand the product model and why it behaves as it does.
- Guides — complete responder and administrator tasks.
- Integrations — connect alert sources and collaboration systems.
- Operate — deploy, secure, scale, back up, and upgrade OpsKnight.
- Reference — exact API, configuration, permission, and runtime contracts.
- Troubleshooting — diagnose symptoms using observable evidence.
- Develop — build, test, and contribute to OpsKnight.