Kubernetes is a core piece in the cloud-native stack. It is a great orchestration platform with a lot of intuitive automation. However, it is quite complex to configure and reach the ideal automation of your containers, application, and infrastructure. It is hard to keep with the new features that are introduced every few weeks. Also, continuous changes that your team might be pushing to your infrastructure and applications compound that complexity. Magalix KubeAdvisor helps you to move much faster with your automation and keep you always on top of it. KubeAdvisor scans your cluster and gives you a detailed report with recommendations to improve it at 10s of dimensions.
Whether you are just getting started or running production-grade Kubernetes clusters, Magalix KubeAdvisor covers a wide variety of checks and tests. For example, having the right observability pipeline requires many components to be installed, configured and maintained. KubeAdvisor will give you a detailed observability analysis and recipes to improve that pipeline. Also, you might be missing some critical configurations to have reliably deployment mechanism of your containers. Magalix KubeAdvisor will provide recommendations to tweak your deployment policies.
KubeAdvisor Covers Key Production Readiness Areas
KubeAdvisor is extensible and we will keep adding advisors. Our initial release covers four main areas:
- Performance by continuously watching throttled containers/apps and recommending improvements.
- Utilization by comparing used resources with the available capacity to reallocate them based on variable workloads.
- Cost Optimization by suggesting changes at the VM level to save money in case of cloud infrastructure or identify the best configurations if you are running Kubernetes on-prem.
- Reliability by observing health probes, deployment configurations, etc.
- Observability through a detailed analysis of the official cloud-native monitoring pipeline.
We keep adding more checks and automation. If you are looking for a particular check and its automation, please drop us a line.
After connecting your cluster, Magalix KubeAdvisor starts analyzing your its structure and different configurations. It generates a report and recommendations to improve the cluster at different dimensions, such as performance, reliability, security, cost, etc.
You can filter by any entity inside your infrastructure and see available recommendations to improve its performance, observability, efficiency, etc.
Anatomy of a Recommendation
Each recommendation covers a certain improvement that you can apply to your cluster or container. 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. The above sample snapshot, for example, proposes a change to nodes types, sizes, and/or numbers.
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 Magalix agent was able to observe to generate this recommendation. Think of as an explainer to the collected evidence. In the above example, this section explains what Magalix agent was able to observe about the current capacity and compares it to the current cluster capacity. It provides a detailed analysis of the current and suggested utilization, capacity, VM types, and different billing plans.
This section shows the relevant metric or meta-data that Magalix agent collected from your cluster. The above snapshot shows the aggregated usage, allocated, and currently available of CPU and memory. It provides a visual comparison to help you decide if you are under or over-utilizing the capacity that you allocated for 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 generic instructions to fix it.
Resources to Learn More
This section provides links from the community and verified blogs to learn about the target category of issues reported in any given recommendation.
You see in this section how many times the issue was detected and a recommendation provided by Magalix KubeAdvisor.
Run KubeAdvisor Recommendations on Autopilot
Magalix currently automates performance and utilization management. When you turn-on the Autopilot, Magalix agent continuously scale applications for performance, utilization, and reliability. Some areas require application-specific logic that we only point out for you, such as implementing the Liveness probes.
Have Your Cluster Check For FREE!
If you haven’t connected your Kubernetes yet, Register and connect your existing cluster. Within a few minutes, cluster metrics will start flowing and our AI engine will engage shortly analyzing your containers and infrastructure.