Kubernetes monitoring requires that you’ve already come to terms with how cloud orchestration works in principle. Specifically, it presumes you’ve already moved past the initial phases of containerization, and are now seeking solutions that fit your operation(s). Consider our blog about ‘Container Orchestration’ for a more general approach to the concept.
If your answers aren’t covered in any of this content, you can always contact an expert about Kubernetes Monitoring by clicking the button below.
What Should I Monitor in Kubernetes?
Before going into the means, it’s important to understand how Kubernetes monitoring is useful. These are the main attributes you would track in any container orchestration platform, though each can have several subcategories depending on the depth of your analytics strategy.
- Overall Health – This generally refers to the health of each cluster, or set of nodes, but can scale upward or downward according to company needs. It encompasses a combination of the lesser metrics that follow.
- Latency – This is a measure of how quickly are API requests being resolved.
- CPU Usage – Every software-related metric must consider how much load is placed on the associated CPUs, whether hardware or virtual.
- Disk Usage – Running out of virtual disk space is no big deal if you have automations in place that create a new one. Running out of physical hard disk space can lead to write errors and loss of critical data.
- Pod Deployment – Kubernetes pods are just their name for their smallest deployed unit, which can consist of one container or several. Since the orchestration tool is predicated on how well it manages containers, managing pods is essentially just a translated metric for its own math.
Kubernetes Monitoring Tools
The following is a list of useful tools for Kubernetes monitoring. Note that this list isn’t comprehensive, as new tools pop up all the time and improvements are made to existing platforms. Still, these are a good starting point to help you along the road to figuring out your own Kubernetes monitoring strategy:
- Kubernetes Dashboard – The built-in dashboard might lack some of the features of third-party applications, but it has the distinct advantage of being made by the same developers who make the platform.
- Prometheus – While free and open source, Prometheus was first created for SoundCloud, which means it should handle any volume of nodes from all corners of the map.
- Dynatrace – This solution really focuses on taking Kubernetes monitoring into the future, with enterprise-level features and myriad ways to sort big data.
- cAdvisor – A lightweight and readily-available add-on to any Kubernetes setup, cAdvisor has become one of the go-tos for transitional analytics users.
- Elastic Stack – This group of applications includes Elasticsearch, Kibana, and Logstash, which leverage high-level search capability for finding container orchestration data and presenting it plainly.
Expert Help on Kubernetes Monitoring
Foghorn Consulting has several services that include Kubernetes monitoring. We recommend starting with the first, or contacting us directly if you’d like to speed things along with direct questions.