Skip to main content

Create Jira bugs from failed Playwright tests

Filing the ticket is the step that gets skipped. It happens after the run, when whoever noticed the failure is already three tabs away. Making it part of the workflow means the ticket exists before anyone has to remember to create it.

What you need​

Connect Jira​

The Jira integration uses OAuth and looks up your projects once connected, so the node configuration offers a project picker rather than asking for IDs. It supports both create and update actions.

Build the workflow​

Environment ──sequential──▶ Playwright ──failure──▶ Jira (create)

Draw an On Failure connection from the Playwright node to the Jira node. The ticket is created only when the run fails.

Fill the fields with run context​

A ticket that says "tests failed" wastes the automation. Configure the Jira node so the issue carries enough to triage without opening Playrunner:

  • Summary — the workflow and branch, so duplicates are recognisable at a glance.
  • Description — which specs failed, and a link back to the run for the Playwright report, screenshots, and traces.
  • Project, issue type, and any required custom fields — set once on the node.

Avoiding duplicate tickets​

This is the part worth thinking about before you turn it on. A nightly regression suite that has been failing for a week will file seven tickets if you let it.

Options, roughly in order of effort:

  • File from a narrow workflow. Point the Jira edge at a smoke or critical path suite rather than the full regression run, so tickets mean something.
  • Use the update action. The Jira node supports updating as well as creating, so a follow-up node can append to an existing issue rather than opening another.
  • Put a human in the loop. Send the Slack alert automatically and keep ticket creation on a workflow someone triggers deliberately.
  • Triage first. Route the failure through AI analysis and file the summary, so a repeated failure at least reads as the same problem.

There is no built-in deduplication today. Design the workflow so that a ticket per failure is the behaviour you actually want.

Alerting and filing at once​

Two failure edges out of the same Playwright node both run on the same failure:

                   ┌──failure──▶ Jira   (create bug)
Playwright ────────┤
                   └──failure──▶ Slack  (#qa-alerts)

Next steps​