Are there currently any guarantees about webhook delivery (e.g. keep trying until delivered “at least once”)? Does the system currently retry?
Is there any interest in making this more robust, like a mechanism for reporting success back to Beeminder, and perhaps signing the webhook contents?
Are you talking about the zeno webhooks? If we don’t get a 200ish response code when we dispatch a zeno webhook, we notify the user that the zeno webhook failed, and we include the data we were trying to send.
Yes, I was! No, I wasn’t! I forgot that there’s two kinds; I was talking about the PESOS webhook.
Sorry if I’m missing something obvious, but how does the notification happen exactly; e-mail?
I didn’t immediately find this related thread Webhook/Callback documentation?, which says that nothing happens when they fail.
Zeno webhooks send that notification via email.
I don’t think PESOS webhooks retry or notify if they fail, but it does look like the failure is logged internally.
To answer earlier questions–I think making them more robust would be fine, but I’m not a priority-driver; that’s a @dreev and @bee thing.
(What’s the motivation behind signing the contents? Making sure the payload was delivered intact?)
Since you’re using e-mail, I don’t think you saw my edit (sorry). I misspoke and wasn’t talking about the zeno webhooks…
Hey Adam, just a friendly ping because we got our wires crossed yesterday…
Webhook systems typically retry until successful delivery. Adding success reporting and content signing could boost reliability and security.