I want to use a WebHook from an automated build repository to trigger a Jenkins job, and because Jenkins’ needs are simple (just a POST, any POST) it basically works. But Docker Hub expects a callback request from the WebHook recipient so that it can notify subsequent URLs in its WebHook chain, and there’s no way to retrieve the necessary data from the WebHook POST body (Jenkins ignores it) to issue such a request from the Jenkins job. Consequently, Docker Hub is left hanging. This is obviously distasteful, but is it problematic? Is there a timeout, or would incomplete invocations just stack up? Or worse, would Docker Hub block (or drop) a new WebHook event if it’s still waiting for callback notification from a prior one?
There isn’t any harm or any policies for timeout. Currently, it just stays in pending. However, in near future, we would like a way to have them change state from pending. Even then, it should not do any harm to your Jenkins job.
Thank you so much for the rapid and informative response, @sbasyal ! It sounds like I’d just have bunch of these unfinished chains piling up after a while. Feature request: an ability to delete those bad boys.
For now you can manually cancel it, using the same user (perhaps jenkins user) that created the webhook from Web UI. But definitely it would be better to have a feature to delete all waiting webhooks at once (and from any user in that organization).