Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
1.1-beta-2
-
None
Description
add a throttle on messages, particularly for errors - don't send a message that is identical to one sent in the last X hours.
Trying to avoid the mass-spamming when a service goes down (but ensure some notification does go out). As soon as the error state is reset to good, another message can be sent, but if it is already in error it shouldn't (currently, it does it every time, and alwaysSend is not granular enough).
On the flipside - I sometimes don't see errors ever get notified (or only once, and then they are stuck in error). We should re-send if something remains in error after a certain duration.
Attachments
Issue Links
- is duplicated by
-
CONTINUUM-1656 If subversion is not available continuum will continue to send an ERROR message until it is available
- Closed