When a webhook fails, to determine the exact reason for failure, you have to change the compute engine log level to DEBUG and then run a new analysis and then download the logs and look for the webhook failure line. We see this happen often as we test and help customers work with webhooks, especially because of certificate errors.
1.) Maybe bubble up the webhook error to the webhook UI. At the very least log the webhook error at INFO level… it is an error after all, so seems like it should be higher than DEBUG level? This would at least make it so that the log level doesn’t need changed and an analysis rerun, and also avoid loss of information if a system restart knocks DEBUG back to INFO. Surfacing details at the webhook UI would be even better though.
2.) Provide more detail about certificate errors. Make it more clear why the certificate is failing. Not in trust store, host name miss-match, etc.?
3.) Provide a “Test Webhook” button to see if the target URL is good to receive data (maybe a server was down, network error, or something). This could be resending a previously failed payload, or sending a test or no data payload.