Hi,
I am using the Prometheus Python client library to instrument my application and send metrics to AMP with the help of AWS Distro Otel (ADOT) collector in between. I am creating a custom timestamp for my metric which is a few minutes older than the current time.
For example, let's say the current time is 12:00:00 and I had to report a metric value 100 that occurred at 11:50:00. I am creating a metric with value 100 and custom timestamp 11:50:00.000 (milliseconds precision) and sending it to AMP using ADOT in between.
However, on the AMP side, I can see the metric value 100 being reported from 11:50:00 to 11:55:00. That is the value 100 is being reported up to +5 minutes from the timestamp that I had intended it to be actually reported.
I believe the AMP should not be filling such wrong values on its own.
This has turned to be a major blocker for us and any immediate help or support on this would be really appreciated.
