Closed kazysgurskas closed 5 years ago
Is this still an issue? I've reviewed how this works and there's nothing obvious around an order of magnitude transcription error, so not sure why you'd see this. We take the sum of all the samples since we last checked, and then divide by the interval since we last checked, which gives us an average over the period. If this still an issue, can you drop detailed logs into the issue and this will show the interval being used?
Do not have a way to replicate this anymore, sorry :) Never got it to work, so quit trying. Thanks for the reply, though!
Running autoscaling 9.5.5. It seems that StreamMonitor collects wrong data from CloudWatch. I'm seeing roughly 10x lower numbers in the logs:
In CloudWatch dashboard under same timestamp I'm seeing 250 records and ~1.6 MB PUT capacity usage. This causes the application to scaledown aggressively.