Releases: coroot/coroot
0.12.4
0.12.3
0.12.2
0.12.1
0.12.0
Deployment Tracking
Coroot now automatically discovers and monitors every application rollout in your Kubernetes cluster.
This feature allows you to stay informed about any changes in your infrastructure.
How it works
Coroot 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:
0.11.1
0.11.0
Memory leak detection
Coroot now monitors the memory usage of each application over time and detects memory leaks before the OOM killer restarts their containers.
The default threshold for detecting a significant increase in memory usage is set at 10MB per hour, but it can be easily adjusted for any application or an entire project.