Uploaded image for project: 'Jira Data Center'
  1. Jira Data Center
  2. JRASERVER-60518

WebHooks not firing when the event triggered is Workflow post function

XMLWordPrintable

      NOTE: This bug report is for JIRA Server. Using JIRA Cloud? See the corresponding bug report.

      We need to transfer the issue status to the external system, so we've created a workflow with the "trigger a webhook post function." The webhook should only work with transitions, so no issue events are selected.

      It works unstable - sometimes webhook fires immediately after the transition and all the required info is posted in the external system.
      Sometimes it works with the time delay (various time intervals).
      Sometimes it doesn't fire at all.

      It is only logged when the webhook was sent successfully.
      We've tried all the options from https://jira.atlassian.com/browse/JRA-41388, but all connections are ok.

      Are there any other suggestions of the possible reason for the problem?

            [JRASERVER-60518] WebHooks not firing when the event triggered is Workflow post function

            Pierre Kroll added a comment -

            Is there any update to this feature?

            Pierre Kroll added a comment - Is there any update to this feature?

            Janne Korkkula added a comment - - edited

            Scott, Re. MS Teams, we're facing the bad request issue too, but you can't just follow some link from the logs and assume your request would be anything like the one Jira has sent to the endpoint. The error means that the webhook has triggered, so this is totally unrelated to this specific issue.

            Nevertheless...

            Point a dummy webhook to any plain old HTTP server and take a tcpdump of traffic to that server. Following the HTTP stream, you'll get the actual POST request. You can then copy the JSON data to a file and feed it to the real webhook endpoint with curl using the exact same headers Jira did. Alas, at least for us connectors.msteams-atlassian.com just throws the 400 with no explanation whatsoever, so it was an interesting but futile exercise...

            curl --trace-ascii - \
  -H 'Expect:' \
  -H 'Content-Type: application/json; charset=UTF-8' \
  -H 'Accept: */*' \
  -H 'Via: 1.1 localhost (Apache-HttpClient/4.5.5 (cache))' \
  -H 'Connection: Keep-Alive' \
  -H 'User-Agent: Atlassian HttpClient 1.0.1 / JIRA-7.13.0 (713000) / Default' \
  -H 'Host: connectors.msteams-atlassian.com' \
  --data-binary @/tmp/jsonfile \
  'https://connectors.msteams-atlassian.com/JiraServer/some-long-hex-stuff@more-hex/even-more-hex/hex-anyone/hex-hex-hex-hex?user_id=user%40domain&user_key=user%40domain' 

...

== Info: upload completely sent off: 9088 out of 9088 bytes
<= Recv header, 26 bytes (0x1a)
0000: HTTP/1.1 400 Bad Request
<= Recv header, 20 bytes (0x14)
0000: Content-Length: 29
<= Recv header, 41 bytes (0x29)
0000: Content-Type: text/plain; charset=utf-8
<= Recv header, 28 bytes (0x1c)
0000: Server: Microsoft-IIS/10.0
<= Recv header, 23 bytes (0x17)
0000: X-Powered-By: ASP.NET
<= Recv header, 37 bytes (0x25)
0000: Date: Wed, 13 Feb 2019 09:24:08 GMT
<= Recv header, 2 bytes (0x2)
0000: 
<= Recv data, 29 bytes (0x1d)
0000: Failed to handle notification
Failed to handle notification== Info: Connection #0 to host connectors.msteams-atlassian.com left intact

            PS. I think you just exposed the keys to your Teams.

            Janne Korkkula added a comment - - edited Scott, Re. MS Teams, we're facing the bad request issue too, but you can't just follow some link from the logs and assume your request would be anything like the one Jira has sent to the endpoint. The error means that the webhook has triggered, so this is totally unrelated to this specific issue. Nevertheless... Point a dummy webhook to any plain old HTTP server and take a tcpdump of traffic to that server. Following the HTTP stream, you'll get the actual POST request. You can then copy the JSON data to a file and feed it to the real webhook endpoint with curl using the exact same headers Jira did. Alas, at least for us connectors.msteams-atlassian.com just throws the 400 with no explanation whatsoever, so it was an interesting but futile exercise... curl --trace-ascii - \ -H 'Expect:' \ -H 'Content-Type: application/json; charset=UTF-8' \ -H 'Accept: */*' \ -H 'Via: 1.1 localhost (Apache-HttpClient/4.5.5 (cache))' \ -H 'Connection: Keep-Alive' \ -H 'User-Agent: Atlassian HttpClient 1.0.1 / JIRA-7.13.0 (713000) / Default' \ -H 'Host: connectors.msteams-atlassian.com' \ --data-binary @/tmp/jsonfile \ 'https: //connectors.msteams-atlassian.com/JiraServer/some- long -hex-stuff@more-hex/even-more-hex/hex-anyone/hex-hex-hex-hex?user_id=user%40domain&user_key=user%40domain' ... == Info: upload completely sent off: 9088 out of 9088 bytes <= Recv header, 26 bytes (0x1a) 0000: HTTP/1.1 400 Bad Request <= Recv header, 20 bytes (0x14) 0000: Content-Length: 29 <= Recv header, 41 bytes (0x29) 0000: Content-Type: text/plain; charset=utf-8 <= Recv header, 28 bytes (0x1c) 0000: Server: Microsoft-IIS/10.0 <= Recv header, 23 bytes (0x17) 0000: X-Powered-By: ASP.NET <= Recv header, 37 bytes (0x25) 0000: Date: Wed, 13 Feb 2019 09:24:08 GMT <= Recv header, 2 bytes (0x2) 0000: <= Recv data, 29 bytes (0x1d) 0000: Failed to handle notification Failed to handle notification== Info: Connection #0 to host connectors.msteams-atlassian.com left intact PS. I think you just exposed the keys to your Teams.

            Scott Boisvert added a comment - - edited

            Similar issue here, trying to post to MS Teams via the web hook. Works fine if the web hook uses Issue Updated. However, when clearing Issue Updated and just putting the web hook trigger on the transition, it does not work. Monitoring the logs I see the following Warning posted:

            2019-02-12 12:55:08,440 httpclient-callbacks:thread-58 WARN anonymous [c.a.webhooks.plugin.PublishTaskFactoryImpl$PublishTaskImpl] Client error - 400 when posting to web hook at .......................

            When following the link, it appears to be doing a GET which is not allowed:

            Invalid webhook request - GET not supported

            Scott Boisvert added a comment - - edited Similar issue here, trying to post to MS Teams via the web hook. Works fine if the web hook uses Issue Updated. However, when clearing Issue Updated and just putting the web hook trigger on the transition, it does not work. Monitoring the logs I see the following Warning posted: 2019-02-12 12:55:08,440 httpclient-callbacks:thread-58 WARN anonymous [c.a.webhooks.plugin.PublishTaskFactoryImpl$PublishTaskImpl] Client error - 400 when posting to web hook at ....................... When following the link, it appears to be doing a GET which is not allowed: Invalid webhook request - GET not supported

            Andreas Krupp added a comment - - edited

            Any update on this issue? We are trying to implement an interface based on webhooks and looking at this issue, this does not seem a reliable solution. When will this be fixed?

            Jira Version 7.9.2 impacted.

            Andreas Krupp added a comment - - edited Any update on this issue? We are trying to implement an interface based on webhooks and looking at this issue, this does not seem a reliable solution. When will this be fixed? Jira Version 7.9.2 impacted.

            Janne Korkkula added a comment - - edited

            This documented but 100% broken feature inhibits sending issue-closing transitions back to an external system, which creates new issues in Jira, rendering the whole integration attempt mostly useless.

            Two and a half years and many, many releases since this bug was reported, what's going on?

            We need this to work (or to have an Issue Closed event type added, but this needs to be fixed anyway).

             

            After upgrading to 7.13.0, connection errors have been logged to catalina.out, so in our case the workflow post function webhooks do get triggered.

            Janne Korkkula added a comment - - edited This documented but 100% broken feature inhibits sending issue-closing transitions back to an external system, which creates new issues in Jira, rendering the whole integration attempt mostly useless. Two and a half years and many, many releases since this bug was reported, what's going on? We need this to work (or to have an Issue Closed event type added, but this needs to be fixed anyway).   After upgrading to 7.13.0, connection errors have been logged to catalina.out, so in our case the workflow post function webhooks do get triggered.

            Darin Hafer added a comment -

            Jira Webhook for Glip sends notifications on issue_updated as expected, but no notification is sent when the issue is transitioned to DONE. When you close an issue, isn't that considered an issue_updated action? 

            I never receive a notification on DONE, so I trigger the webhook in the post-function of the transition to DONE, but it never triggers.

            For how long will this issue be ignored?

            Darin Hafer added a comment - Jira Webhook for Glip sends notifications on issue_updated as expected, but no notification is sent when the issue is transitioned to DONE. When you close an issue, isn't that considered an issue_updated action?  I never receive a notification on DONE, so I trigger the webhook in the post-function of the transition to DONE, but it never triggers. For how long will this issue be ignored?

            Наталия Ицкович added a comment -

            We're using the following workaround - when the external system accepts the new info, it sends back to JIRA  the text fileld SUCCESS. and we empty this field at every transition right before triggering webhook. So if webhook worked SUCCESSfully, this field is never empty.

            Afterwards in all the cases when the success-field is empty (JQL filter) , the external system 'asks' JIRA to send the tickets again - and we do this check by job every 10 minutes.

             

            So we have 2 integration flows from jira.

            Наталия Ицкович added a comment - We're using the following workaround - when the external system accepts the new info, it sends back to JIRA  the text fileld SUCCESS. and we empty this field at every transition right before triggering webhook. So if webhook worked SUCCESSfully, this field is never empty. Afterwards in all the cases when the success-field is empty (JQL filter) , the external system 'asks' JIRA to send the tickets again - and we do this check by job every 10 minutes.   So we have 2 integration flows from jira.

            Shankar Esakki added a comment -

            Hi Team, 

            We are facing the same kind of issue, time interval for the data is not updating(WebHook) is "12:00 Am IST to 5:00 AM IST" 

            Please check and resolve the problem immediately. 

            Regards,

            Shankar. E

            91-9884957983

            Shankar Esakki added a comment - Hi Team,  We are facing the same kind of issue, time interval for the data is not updating(WebHook) is "12:00 Am IST to 5:00 AM IST"  Please check and resolve the problem immediately.  Regards, Shankar. E 91-9884957983

            Paul Morris added a comment -

            Hi, we are experiencing this issue too, has there been any update?

            Thanks,

            Paul

            Paul Morris added a comment - Hi, we are experiencing this issue too, has there been any update? Thanks, Paul

              Unassigned Unassigned
              d1ea29b53423 Наталия Ицкович
              Affected customers:
              52 This affects my team
              Watchers:
              42 Start watching this issue

                Created:
                Updated: