From the community!
- fix broken link to CONTRIBUTING.md - kubecost/cost-analyzer#8
- Use consistent tags in deployment utilization dashboard - #1043
API
- CSV download for Asset and Allocation APIs
Configuration
- Add generic cost-analyzer network policy - #1013
- Added a shared overhead costs parameter to the helm config - #987
Cost
- Fixed Azure CSV parsing issue
- Fixed edge case in which idle would not be shared between allocations - kubecost/cost-model#883
- Fix for AWS provider node pricing estimates used in cluster sizing. Issue did not affect features other than cluster sizing.
- Add Start, End, and Minutes to AssetSetRange - kubecost/cost-model#905
- Fix parsing of configured SharedOverhead in AggregateCostModel to gracefully handle empty string - kubecost/cost-model#920
- Added guard in spot checklist to prevent edge case cost-model panic due to uncommon/unexpected nil node
Frontend
- Fixed issue with loading and saving of allocation reports with idle allocation
- Fixed issue where Cluster Sizing savings showed "Infinity%"
- Add cloud diagnostics and update ooc check on assets page
- First version of Cloud Connection Status
- Prometheus metrics diagnostic optimizations - kubecost/cost-model#913
- Tooltip to explain infinite efficiency
- Fixed navigation bug triggered by long cluster names
- Fixed issue where long requests on Reports would overwrite each other
- Fixed cadvisor diagnostics
- Fixed issue where, on initial startup, the cluster info could be empty if cost-model down for 5+ minutes - kubecost/cost-model#917
- Limit one filter per property on Request Sizing page
- Prometheus Request queue diagnostics - kubecost/cost-model#889
- Correction to health calculation on the frontend
- Fixes to filter behavior when using drilldown and filters together
Performance/Scalability
- Migrated container network queries to new cadvisor data, including optimizations - kubecost/cost-model#907
- Quieting of the RequestSizing logs
- Binary format optimizations
- ETL repair/rebuild improvements, including memory and speed optimizations, in addition to status improvements