Version 4 introduces significant updates of Densify Container Data Collection. It resides in a new Github repository, which replaces the versions 1-3 repository, to be deprecated on June 30th, 2024.
While the core PromQL queries remain the same, with the exception of some improvements and bug fixes, the container data collection has been re-architected to enable new features and improve performance.
Version 4 allows you to run a single job to collect data for multiple Kubernetes clusters from a single Prometheus server / observability platform. Using Prometheus labels, the job determines which data belongs to which cluster.
Due to Prometheus' data retention and long-term storage limitations, many customers are now using observability platforms.
Version 4 collects data from external Prometheus servers (Prometheus running outside the Kubernetes cluster where the data collection runs), as well as external observability platforms that implement the Prometheus API. Specific authentication mechanisms are included to facilitate the access control these platforms require.
- Adds HTTP retries to calls to the Prometheus API (handles observability platform rate limiting)
- Handles popular relabel configs of Node Exporter
- Fixes outdated Node Exporter metrics
- Uses improved Horizontal Pod Autoscaler metrics (autoscaling v2)
- Fixes issues with Kubernetes Resource Quota metrics collection
- Handles various changes and additions to kube-state-metrics
- Adds collection of cAdvisor's
container_memory_working_set_bytes - Upgraded to Go 1.22
- Updated documentation and examples