CPU not showing in OS/CPU Utilization Dashboard

Percona Monitoring and Management (PMM) PMM 2.x
1

OS/CPU Utilization Dashboard is missing CPU utilization. It seems timeout is happening based on the pmm-agent.log. Version is 2.32.0.

ERRO[2022-11-29T19:33:37.950+00:00] time=“2022-11-29T19:33:37Z” level=error msg=“POST https://logs.us-east-1.amazonaws.com/ → context deadline exceeded (579.170685ms)” component=transport source=“transport.go:52” agentID=pmm-server/rds component=agent-process type=rds_exporter
ERRO[2022-11-29T19:33:37.950+00:00] time=“2022-11-29T19:33:37Z” level=error msg=“Failed to filter log events: RequestCanceled: request context canceled\ncaused by: context deadline exceeded.” component=enhanced source=“scraper.go:142” agentID=pmm-server/rds component=agent-process type=rds_exporter
ERRO[2022-11-29T19:33:38.894+00:00] time=“2022-11-29T19:33:38Z” level=error msg=“POST https://logs.us-east-1.amazonaws.com/ → context deadline exceeded (1.544161166s)” component=transport source=“transport.go:52” agentID=pmm-server/rds component=agent-process type=rds_exporter
ERRO[2022-11-29T19:33:38.894+00:00] time=“2022-11-29T19:33:38Z” level=error msg=“Failed to filter log events: RequestCanceled: request context canceled\ncaused by: context deadline exceeded.” component=enhanced source=“scraper.go:142” agentID=pmm-server/rds component=agent-process type=rds_exporter

Not sure what the issue is and if hitting some kind of quota limit. We have 90 plus servers across several IAM Roles. The main one with the issue has around 45 RDS Instances. All the IAM Roles are setup with the same permissions and been using for years.

1 Like
2

The RDS Instances are RDS MariaDB, RDS MySQL, and Aurora MySQL.

1 Like
3

Did you follow all the steps for using PMM with an IAM role?

1 Like
4

yes, I sure did. We have multiple IAM Users in different PMM environments with no issues. Not sure what the issue is. I checked to make sure each RDS instance was valid and under the correct IAM User.

1 Like
5

Some of the databases have large files on the RDS instance side. Could that cause issues?

1 Like
6

It should not because PMM pulls the CPU info from CloudWatch. Do you see CPU/OS/Disk metrics for these instances in CloudWatch?

1 Like
7

I will make sure Enhanced Monitoring is enabled on all nodes. That did cause an issue in the past.

1 Like