Cost Model
- Add better spot node estimates to azure
- Fix issue where out-of-cluster costs for 2-4 days would occasionally disappear
- Fix issues where repairing large ETL pipelines would sometimes return timeouts to the users
Frontend
- Fix issue where namespace.html recommendations were not consistent
- Add efficiency numbers to csv downloads
- Add asset aggregation by provider/account/providerID to assets.html
- Upgrade jquery
- Fix issue where "other" category would not be added to range charts correctly
- Fix unknown aggregation: null errors on asset page
Alerts:
- Send slack diagnostic alert when a cluster joins or leaves the list of clusters accessible by Thanos
- Spend change alerts
Helm Chart
- Fix issue where large cookies would fail to be passed to the frontend
- Fix issue with awsstore pods crashing
- Fix verbose logs in Network Costs
- Support kubernetes 1.22