In a Kubernetes ecosystem, with multiple clusters deployed, it is not only time consuming, but tedious for operators/developers to debug issues. Application deployment data are stored in various dispersed systems and difficult to navigate, audit and diagnose. K-Atlas solves this by providing a holistic view and query interface.
Additionally, there is no graphical view of the topology, making it hard to visualize the system. K-Atlas solves this by providing a near real-time graphical view of the system.
- Evaluate the configuration of multiple K8s clusters (potentially in multiple regions)
- Compare against baseline and each other, and provide a diff
- Present a global view of an application - across clusters and regions
- What versions of the application are running? Where?
- What are the key objects powering the app? What is their current status? How are they changing?
- Monitor application fleet on a single pane of glass - how many resources are my application consuming? In what regions?
- Diagnose and narrow down application problems
- How many pods are backing a dns domain x.y.abc.com?
- What is the application deployment that has pod IP a.b.c.d?
- Set policies and query for compliance
- Send out alerts when policies are violated and aid in enforcement
- Examples of policies include:
- Does application deployment meet minimum DR requirement?
- Run only approved protocols on approved ports (HTTPS on 443)
- Playback changes that happened on the fleet
- Provide deep visibility into any change to the fleet