Metrics FAQs

Overview

Amazing news! You've onboarded your Kubernetes clusters to Komodor, with this installing the Komodor-agent and the Komodor-Agent-Daemon. Let's look at some questions that almost always come up.

How to know if metrics are enabled for a cluster?

Once the cluster has the Komodor Agent and Komodor-agent-daemon installed, the metrics flag should be enabled by default.

How to know if both elements of the Komodor Agent are installed?

You can ensure this is the case by going to the Services UI and selecting the cluster and looking for komodor. You should see both the agent and the Daemonset, as seen below. Without both elements installed, Metrics will not populate.

How long after installation should there be metrics showing?

There are a couple of types of data we can expect to be shown under Metrics
- Raw Data (Pods/Nodes) should be available a few minutes post-installation.
- Workload metrics (or anything that requires aggregation of data) - this may take up to a couple of hours depending on when the agent was installed. You should expect one data point per hour.

How are the metrics calculated?

We’ve updated the metric we use to calculate memory consumption. Previously, we used memory_usage_bytes, and we’ve now switched to memory_working_set_bytes. This change aligns with standard practices for memory measurement, providing a more accurate reflection of actual memory usage.

You may notice a slight adjustment in displayed memory figures starting November 5th, but rest assured, this is simply a reflection of the updated calculation. No issues to worry about - just an improvement to ensure even better accuracy.