How to get the best out of KubeAdvisor recommendations
After connecting your cluster Magalix KubeAdvisor starts analyzing your cluster's data. It generates a report and recommendations to improve the cluster at different dimensions, such as performance, reliability, security, cost, etc.
A recommendation is how to fix an issue. Each workload gets its unique recommendation to address a specific reported issue.
Anatomy of a Recommendation
Each recommendation covers a certain improvement that you can apply to an entity within your cluster. But all recommendations have the same overall structure.
Each recommendation consists of 7 main sections
It tells you what you need to change to fix a reported issue. In many cases, the required change is specific and quantifiable.
This section tells a quantifiable impact whenever possible. For example, this cost-saving recommendation tells how many yearly savings you will achieve
This section explains what the Magalix agent was able to observe to generate this recommendation. Think of as an explainer to the collected evidence.
This section shows the relevant metric or meta-data that the Magalix agent collected from your cluster.
How to Resolve the Issue
This section explains and links to different resources to resolve the reported issue. Magalix in some cases provides an out of the box automation to resolve the issue. If not automation possible, you will get instructions to fix it.
You see in this section how many times the issue was detected and a recommendation provided by Magalix KubeAdvisor.