Description
We have a SQS queue, we increased its Visibility timeout to 300 sec, previously it was 30 secs. It also has a DLQ setup.
These messages are processed by a Lambda function using a Trigger (Batching Window = 10 sec)
Problem
There are spikes in Approximate Age Of Oldest Message metric around 300-340 sec.
We have checked lambda logs, there are no errors/exceptions and SQS DLQ is also empty which indicates there was no message processing error.
we noticed that, when the visibility timeout was 30sec, the spikes were still there, but smaller, because visibility timeout was only 30 secs.
My theory is that, the spikes will be always be around whatever is the visibility timeout.
Question
What could be the reason behind these spikes?
