Jira

Overview

The Jira integration opens an issue in your project when a monitor goes down. Available on the Enterprise plan.

Note

Only a DOWN transition creates a ticket — up, degraded, and paused deliberately do nothing, so a flapping or recovering monitor never spams the project. Resolving/closing the issue is left to your own workflow. For on-call alerting (create + auto-resolve), use Jira Service Management instead.

Setup

1. Create an API token at id.atlassian.com → API tokens.
2. Note the project key (e.g. OPS) and the issue type you want (e.g. Task, Bug).
3. In OKStatus, go to Alerting → Channels → Add a channel → Jira.
4. Fill in:
   • Base URL — https://your-team.atlassian.net
   • Account email — the Atlassian account the token belongs to
   • API token
   • Project key
   • Issue type (optional, defaults to Task)
   • Reuse & auto-resolve one ticket (optional checkbox — see below)
5. Click Test, then attach the channel to your monitors.

Reuse & auto-resolve one ticket

By default the channel opens a new ticket every time a monitor goes down — which, for a monitor that flaps (down/up repeatedly), can flood your board. Enable Reuse & auto-resolve one ticket to instead keep a single ticket per monitor that tracks its state:

• Down → opens a ticket (or, if one already exists, reopens/comments on it).
• Up → transitions that ticket to a Done status.
• Down again → reopens the same ticket instead of creating a new one.

So a monitor that flaps 30 times produces one ticket with 30 status changes, not 30 tickets.

How it works, and why it’s robust:

• The ticket is found each time by a monitor-specific label (OKStatus never stores the issue key), so the behaviour is stateless.
• Transitions are auto-detected from your project’s workflow (it picks a transition whose target status category is Done to resolve, or a non-Done one to reopen) — no status names to configure, works with custom workflows.
• If an admin deletes the ticket entirely, the next down simply creates a fresh one (the label search finds nothing), and a recovery with no ticket is a safe no-op. Nothing breaks.

Note

Jira’s search index is eventually consistent, so if a monitor flaps down→up within a few seconds, the recovery may not yet find the just-created ticket and could leave it open; the next event reconciles it. Real incidents (minutes or longer) are unaffected.

Notes

• Authentication is Basic (email + API token), over the Jira REST API.
• The created issue’s summary and description name the monitor, URL, region, and detection time.
• Self-hosted Jira (Data Center) works — use its DNS name (a literal private IP is rejected).