Closed kirg closed 7 years ago
There's some missing stuff here. One thing is that we are preventing messages from getting into message cache, but we really ought to expire messages in the cache as well (by acking them). Some logic in messagecache.go would solve this. If you pursue that, be sure that you appropriately expire the smart-retry LKG message.
The 'skip-older' is primarily intended as a mechanism to flush out the backlog quickly. And what you suggest does help enforce 'skip-older' more strictly .. but I am going to leave this out for now; we can perhaps consider this in the future if we find that there are too many messages hanging around in the message cache for unacceptably long.
Coverage decreased (-0.03%) to 68.492% when pulling 3517106ba62c36a494cdb3ce41ab34aec819673d on outputhost-delay-and-skipolder into ec8103f2637ad4749c8880d88fb56b8c953a8af6 on master.