I setup the following flow and webhook with various testing:
How can I diagnose the issue to fix the connection between Webhook and Automation Rule?
Hi @Thomas Behlau ,
Thanks for the post.
With my team, we went through a similar issue with Incoming Webhooks in automation.
Our issue was related to how the Automation incoming webhook secret was being passed. And our solution was to append the Automation secret to the URL and leave Jira’s System Webhook Secret field blank.
This entry in the Atlassian Support blog helped us solve the case: https://support.atlassian.com/cloud-automation/docs/configure-the-incoming-webhook-trigger-in-atlassian-automation/
It provides specific steps to set up an incoming webhook in automation.
I hope this helps, and reach out if you have any further questions.
Cheers,
Fede.
The reason you have nothing to diagnose with: the automation audit log only records requests the webhook endpoint actually accepted. A request that fails token validation is rejected before any rule runs, so it leaves no trace on the Jira side. Silence in the audit log almost always means the URL is wrong, not the rule.
Two things to check, in order:
1. The full trigger URL, token included. The incoming webhook trigger's URL already carries its secret as part of the path (.../automation/webhooks/jira/a/<id>/<token>). Fede's suggestion is the one to test first: if you copied the base URL and put the secret into the system webhook's "Secret" field instead, the request hits an incomplete endpoint and dies unlogged. Jira's Secret field does something entirely different anyway - it HMAC-signs the payload into an X-Hub-Signature header, which automation never reads. Leave it blank and paste the exact URL from the trigger config, same as your working Postman call.
2. The payload shape, once requests start landing. In the trigger settings, check "Execute this automation rule with": the "issues provided in the webhook body" option expects {"issues": ["KEY-1", ...]}, but a system webhook sends a single issue object - so the rule fires with zero issues and your log action never executes, which looks identical to "not working". Set it to "No issues from the webhook" and read the payload with {{webhookData}} smart values ({{webhookData.issue.key}} etc.), or use the run-JQL option if later actions need real issue context.
To pin it down: point the system webhook at webhook.site once more, capture the exact body it sends, then replay that body from Postman at the automation URL. Postman works and the replay carries the system webhook's real payload, so whatever breaks in the replay is your culprit - and it will be one of the two above.
I'm assuming you're routing through a system webhook because Entity Property Set has no native automation trigger - that part is the right architecture, for what it's worth. It's just badly signposted that the two webhook systems' auth models don't line up.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.