Closed colmsnowplow closed 2 years ago
Note: We have yet to see this observation replicated. Leaving this open for another while in case it does creep back up, but this may be a false alarm.
We couldn't replicate this, and we haven't encountered it again so we'll chalk it down to human error somewhere along the line and re-raise it if we encounter it again.
We have noticed an issue whereby stream replicator is reporting correct metrics via statsd (to cloudwatch), but the observer_buffer logging seems to be reporting at least the
maxMsgLatency()
in seconds where it should be miliseconds. (ie 6000 instead of 600,000).Not sure exactly why - needs investigation.