Integrations2 min read

Troubleshoot Jira

Diagnose credentials, project mapping, issue creation, and webhook synchronization.

Connection test fails

Confirm the base URL, credentials, project permission, and encrypted settings. Run the test again only after correcting the visible failure.

Classify the response: DNS/TLS failure is the network boundary; 401 is the account/token pair; 403 is product or project permission; 404 often means a wrong cloud base URL or project; 429 requires backoff. For Jira Cloud, use the site base URL rather than a copied issue URL. Do not put the API token in a URL.

Issue creation fails

Verify the project key and issue type exist and that required Jira fields are available. Inspect the incident operation before retrying.

Check the configured project key and issue-type ID against the same Jira site. Inspect Jira's field error response for required custom fields, invalid option IDs, or unavailable reporter/assignee. Before retrying an ambiguous timeout, search Jira and the OpsKnight link record for the incident key so a successful remote create is not duplicated.

Webhook updates do not appear

Confirm the integration is enabled, the shared secret matches, and Jira is sending a handled event. Check for 429 responses and preserve the Atlassian webhook identifier when diagnosing duplicate or delayed delivery.

Confirm the Jira webhook URL is the current OpsKnight URL, event subscriptions include the documented issue transitions, and the shared secret is the matching revision. Compare the Atlassian webhook identifier and timestamp with ingress and OpsKnight logs. A 2xx with no visible change may be an ignored event or an issue that is not linked to an OpsKnight incident.

Direction and loop checks

Determine whether the missing change is OpsKnight → Jira or Jira → OpsKnight. Check the outbound operation for the former and webhook delivery for the latter. When both directions are enabled, verify origin markers prevent the same update from bouncing repeatedly. Do not fix a loop by disabling all synchronization without preserving the failing event and link evidence.

Verify and escalate

Create a test issue from a non-production incident, change one supported field in Jira, and confirm one correlated update returns to OpsKnight. Preserve incident ID, Jira issue key, operation ID, webhook ID, response status/body, field errors, and timestamps. Redact email addresses and credentials.

Last updated for v2.0.0

Edit this page on GitHub