This seems to happen very intermittently without a rhyme or reason, and we'd obviously love 100% consistency if we possibly can. Subsequent publishes of different content items did trigger the webhook, as well as subsequent publishes of the same content item. Would the Prismic team be able to shed any light on the issue?
This would be for the leading-edge repo, the publish of a content item at ~5:36PM EST did not trigger the webhook, but the 5:44PM and later publishes did successfully trigger the webhook.
Hi Phil, thank you for following up! I did neglect to notice that Webhook logs could be expanded for more information, though it seems only a limited amount of logs are shown and we've already progressed beyond when the publish event would have taken place. On your end, would you be able to see if the webhook fired at that time successfully? It would have been related to an update to document YbJi2RIAAB4ACIiI. What we had noticed is that the webhook had not reached Netlify to trigger a build of the site on the platform. As of yet, subsequent tests since noticing this have not replicated that issue, unfortunately.
There's nothing on our side in terms of anything failing; if you're seeing a 200 success on our side, this would suggest the issue is happening on the client-side (app). Maybe the Netlify support can help in this case.
We're reaching out to Netlify support as well, but if it's simply that they never received the webhook trigger they may have limited avenues to help us. You're saying that you're not seeing anything failing, but is there any way to see if there was a webhook trigger at all on 5:36 PM EST coinciding with the publish of the update to that content item? As far as I can tell from within the Webhook > Logs view in the dashboard we can only see the contents of the last ten triggers, and it isn't possible for us to view a log that would exist for that trigger from our own end.
Yes, during initial troubleshooting for the error a small update was made to one of its fields and then reverted at 6:05 and 6:08 respectively. This was to test whether the edit and publish would trigger the webhook for the problem content item, which it did in both cases. Unfortunately, history functionality seems to be absent in the new editor, and the legacy editor history does not show exact timestamps, but the three edits to the content item made on the 11th can be seen. No screenshot was taken, but at the time it was logged that there was a Last Published at 5:36PM for that content item (which did not trigger a build) and an edit and publish to a separate item (XjyheRMAACIAjaW-) at 5:44PM, which did trigger a build and thus updated the problem content item.
OK, that's very clear, thank you. Unfortunately, there's nothing I can find in the logs related to this at all. So since we can't find any data on this and can't recreate the behaviour, the only thing I can suggest is that if this happens again please inform us right away, don't make any more changes to the document in question and grab any screenshots etc that might be helpful.