Helm Chart
- Support supplying alert configurations from values http://docs.kubecost.com/alerts
- Support imagepullsecrets for pulling kubecost from private repos
App - Support new N2D prices GKE API
- Fix issue where estimates were incorrectly used for historical pricing on asset pipeline rebuild
- Use "node" label instead of overloaded "instance" label for nodename in kubecost-emitted metrics to avoid potential relabel conflicts
- Use prometheus config data for scrape interval lookup-- defaults to 1minute scrapes
- Reduce clean install initial data gathering time
- Add diagnostics for potential relabeling issues
- Add new diagnostics view to /diagnostics.html for missing ETL data: https://github.com/kubecost/docs/blob/master/diagnostics.md
- GKE out-of-cluster assets added to asset ETL
- Use default spot gpu price as estimate on AWS when prices are not available