Why is it taking an excessive amount of time for my messages to reach my endpoint?
Some of the following scenarios might be responsible:
-
A large operation—such as a bulk update—in the system can cause a large volume of messages to be enqueued all at once, which can take some time to process.
-
Long-running calculations or timeline calculations on large projects could be causing a delay in the publication of messages to Event Subscriptions to consume.
-
The subscription might have been disabled.
-
After a 100-message grace period, if a particular URL—which could be associated with one or more subscriptions—fails more than 70% of the time or if the URL fails to deliver after 2000 consecutive attempts, all messages matching subscriptions with that same URL are not attempted for delivery. Instead, those messages are immediately queued for a retry.
Every 10 minutes after a URL is disabled, we attempt to deliver the next message that comes through for processing. If that message succeeds, then we re-enable that URL and subsequently any matching subscriptions. If that message fails to send, then that 10-minute timer resets and we try again after it expires.
This behavior can be perceived as inconsistent or delayed deliveries, but it simply follows our policies for how Event Subscription messages are handled.
-
An Event Subscription URL will be hard disabled if either of the following conditions are met:
- The Subscription URL has failed to deliver for 7 days and has failed at least 2000 consecutive delivery attempts in the last 72 hours.
- The Subscription URL failed to deliver 50,000 consecutive attempts.
-