Releases: Icinga/icinga-kubernetes-web
Icinga for Kubernetes Web v0.2.0
We are excited to announce the release of Icinga for Kubernetes v0.2.0! This update brings new features and improvements to enhance our Kubernetes monitoring solution, making it easier to analyze problems and understand complex hierarchies.
Vision and Roadmap
Icinga for Kubernetes aims to provide a zero-configuration baseline for monitoring Kubernetes clusters. Our goal is to simplify understanding the complete state of clusters, including resources, workloads, relations, and performance. We leverage the Kubernetes API to retrieve information and store it in a database, reducing API load and enabling powerful filtering.
Future plans include integrating metrics, using Icinga Notifications for alerts, and supporting multiple clusters. We welcome your ideas on key metrics and thresholds by opening an issue on our GitHub repository.
What's New in v0.2.0
Refreshed, Streamlined UI
We've implemented a refreshed user interface for more intuitive navigation. The updated design enhances usability and aligns with modern design principles.
Icinga States
The new Icinga States feature provides a clear overview of resource health, categorizing them into healthy, warning, or critical states by correlating container and pod statuses. This simplifies troubleshooting and gives a clearer picture of your cluster's health.
Annotations Synchronization
Annotations are now synchronized and displayed in the web interface, offering more context and filtering possibilities, making it easier to organize and manage resources.
View YAML of Kubernetes Resources
You can now view the YAML configuration of Kubernetes resources directly from the web interface, simplifying inspection and debugging without leaving the monitoring environment.
Metric Charts
Metric Charts allow you to visualize essential metrics like CPU usage and memory consumption right from the dashboard. This helps in quickly identifying trends and potential bottlenecks for proactive resource management.
Container Log Limitation
To optimize performance and prevent excessive log data, we've limited container log length.
Get Started or Upgrade
To upgrade to v0.2.0, update your packages, containers, or Helm charts based on your installation method. Schema changes will be deployed automatically.
For new installations, any Icinga for Kubernetes components can run inside or outside Kubernetes clusters, including the database. Use our Helm charts to deploy a ready-to-use Icinga stack.
Icinga for Kubernetes Web version 0.1.0
v0.1.0 Version 0.1.0