Deployment tracking
Coroot automatically discovers and monitors every application rollout in your Kubernetes cluster.
This inspection uses metrics gathered by kube-state-metrics
to detect that an application (Kubernetes Deployment) has been updated.
Using Kubernetes metadata to detect rollouts has two main advantages.
Firstly, this is the most up-to-date information about the state of applications.
Secondly, this approach eliminates the need to integrate Coroot with CI/CD tools.
Upon identifying a new rollout, Coroot immediately starts monitoring its status. The possible states of a rollout are as follows:
A rollout is considered "Stuck" if it lasts for more than 3 minutes (the threshold can be easily adjusted for any application or an entire project). This can occur due to issues related to container images or if a new pod cannot be scheduled on a cluster node.
Coroot audits each application 30 minutes after its new version is deployed to evaluate its performance in comparison to the previous version, using the following criteria:
- SLOs (Availability & Latency)
- The number of errors in the application logs
- Container restarts
- CPU consumption
- Memory leaks
Notifications of all notable changes are sent to your Slack workspace to ensure that you don't miss even the slightest performance degradation.
You can turn off such notifications for any particular application category: