Open
Description
It happened to me while troubleshooting two different servers, that the memory graph was not truly reflecting that server was out of memory. @kochetovd mentioned that
The last RAM usage metric value was 15-53 and equal to 15 GB. Then the agent stopped sending data. However, the dashboard shows the average values for 10 minutes.
This was misleading as the host was running out of memory and mysqld
was being OOM-killed.
My proposal is for the graphs to show a gap or some sort of visual warning to tell the user that the agent was unresponsive or that the data sample was not retrieved thus there might be missing values
Metadata
Metadata
Assignees
Labels
No labels