I am trying to setup monitoring of AWS EC2 instances. The task is quite simple, I want to monitor CPU, Disk and Memory usage. Out of the box the AWS template provides CPU usage metrics, but Disk and Memory usage is not available. We have created scripts posting Disk and Memory usage data to CloudWatch as custom metrics. When adding custom metric in the AWS probe, I encounter several issues. I am using AC to configure the AWS probe. UIM ver. 8.5.1, AWS ver. 5.34.
1. Profiles for custom metrics must be added one at a time, and the AWS configuration must be saved and reloaded for each metric added. This is not a good way of adding monitoring profiles in an environment that are changing all the time. I am also unable to find a way to edit the profiles after they are saved.
2. Custom metrics are linked to the IP-address of the robot running the AWS probe, and are not linked to the AWS instances.
Is this how the AWS probe is supposed to work? Do I need to add custom metrics to get used Disk space and Memory usage? These are important metrics to monitor and generate alarms on.
What are the alternatives? Can I deploy probes to the AWS machines and monitor with cdm probe? Do I then need a hub in the AWS environment, or can the AWS machines be connected to a local hub? Any documents on best practices?