Skip to content
Compare
Choose a tag to compare
  • Fix panic in certain multicluster setups
  • Fix panic that occurred when modifying an alerts configmap without restarting the pods
e3373b7
Compare
Choose a tag to compare

From the community!

  • Default support for all Ingress controllers by defaulting Ingress pathType to ImplementationSpecific - #1128 - Thanks @mgoodness!
  • Update Prometheus Ingress logic to accommodate v1 values - #1127 - Thanks @korjek!
  • Comment out Ingress className to avoid conflicts with annotated class names - #1132 - Thanks @gracedo!
  • Support for custom priorityClass in helm chart values - #1097 - Thanks @AshutoshNirkhe!

Configuration/Architecture

  • Ability to disable cost-analyzer-server to support lighter deployments, where applicable.
  • Fix for use of gcpSecretName in helm chart - #1159

Cost

  • Ability to exclude upfront AWS RI and Savings plan costs from adjustments and cloud assets
  • Rate Card API support for Azure GovCloud and Azure China - kubecost/cost-model#994
  • Resolved issue where custom pricing was not being applied to assets - kubecost/cost-model#989
  • Resolved issue where GPU costs incorrectly report $0 - kubecost/cost-model#992
  • Support for additional Azure region pricing (norwayeast) - kubecost/cost-model#996
  • Update to improve reconciliation of load balancer costs
  • Support for disks provisioned by AmazonFSx

Frontend

  • Improved performance of Assets, Allocations, and Request Sizing pages!
  • New Overview page with performance and usability improvements!
  • New Savings page with much improved user experience!
  • Improved performance by avoiding queuing of long-running queries - kubecost/cost-analyzer-frontend#424
  • Improved readability of lower-level drill down information
  • Improved usability of diagnostics
  • Support for shared overhead and shared namespaces in Allocations and saved reports
  • Fix an issue with saved reports not saving
  • Fix an issue with savings not loading on certain on-premise clusters
  • Shows efficiency for rows that are not idle or UnmountedPVs.
  • Show efficiency when aggregating by deployment and daemonset

Performance

  • Improved speed performance of Cloud Asset and Reconciliation processes
  • Ability to tune cloud asset build time performance for extremely large CURs
  • ETL disk usage reductions

Grafana

  • Update to v8.3.2 for CVE-2021-43798 fix
cb836cf
Compare
Choose a tag to compare
  • Fix issue from rc-0 where historical data was not reconciling with cloud data
a5497b3
Compare
Choose a tag to compare

Frontend

  • Fix an issue with saved reports not saving
  • Fix an issue with savings not loading on certain on-premise clusters

Grafana

  • Update to v8.3.2 for CVE-2021-43798 fix
4ae5d82
Compare
Choose a tag to compare

From the community!

  • Default support for all Ingress controllers by defaulting Ingress pathType to ImplementationSpecific - #1128 - Thanks @mgoodness!
  • Update Prometheus Ingress logic to accommodate v1 values - #1127 - Thanks @korjek!
  • Comment out Ingress className to avoid conflicts with annotated class names - #1132 - Thanks @gracedo!
  • Support for custom priorityClass in helm chart values - #1097 - Thanks @AshutoshNirkhe!

Configuration/Architecture

  • Ability to disable cost-analyzer-server to support lighter deployments, where applicable.
  • Fix for use of gcpSecretName in helm chart - #1159

Cost

  • Ability to exclude upfront AWS RI and Savings plan costs from adjustments and cloud assets
  • Rate Card API support for Azure GovCloud and Azure China - kubecost/cost-model#994
  • Resolved issue where custom pricing was not being applied to assets - kubecost/cost-model#989
  • Resolved issue where GPU costs incorrectly report $0 - kubecost/cost-model#992
  • Support for additional Azure region pricing (norwayeast) - kubecost/cost-model#996
  • Update to improve reconciliation of load balancer costs
  • Support for disks provisioned by AmazonFSx

Frontend

  • Improved performance of Assets, Allocations, and Request Sizing pages!
  • New Overview page with performance and usability improvements!
  • New Savings page with much improved user experience!
  • Improved performance by avoiding queuing of long-running queries - kubecost/cost-analyzer-frontend#424
  • Improved readability of lower-level drill down information
  • Improved usability of diagnostics
  • Support for shared overhead and shared namespaces in Allocations and saved reports

Performance

  • Improved speed performance of Cloud Asset and Reconciliation processes
  • Ability to tune cloud asset build time performance for extremely large CURs
  • ETL disk usage reductions
47ac2b6
Compare
Choose a tag to compare
  • Fix issue with GPU costs not reporting correctly to prometheus
Compare
Choose a tag to compare

From the community!

  • Added support for ingressClassName - #1108, #1122
  • Code improvements for Ingress API version processing - #1090

Configuration/Architecture

  • Code refactoring to facilitate more robust deployments - kubecost/cost-model#946
  • Option to reduce duplication of certain KSM V1 metrics while running KSM V2 - #1106
  • Addressed a number of CVEs identified by Twistlock - #1120
  • Addressed issue deploying images for amd64 and arm64 architectures - #1115

Cost

  • Reduce confusion by ensuring a consistent timezone across all cost report providers - #1081
  • Eliminate emission of spurious price values - kubecost/cost-model#935
  • Allow AWS RI offerings to be compared against On-Demand counterparts - kubecost/cost-model#949
  • Improvements to Provider, Project ID, and Account ID asset filtering, including more accurate asset pricing
  • Add support for GCP regional disks for added pricing accuracy - kubecost/cost-model#959

Documentation

Frontend

  • Spot cluster sizing UI
  • AWS Reserved Instance pricing
  • New Alerts, including UI improvements and restart tolerance
  • Cost details now visible at lowest level drill down for Assets and Allocations
  • Numerous improvements to Spot UI, including better messaging, improved layout/styling, and a recommendation command generator (GKE now, others to come)
  • Overall improvements to the pricing source diagnostics, including the addition of Azure Rate Card API connection status - kubecost/cost-model#952
  • BasicAuth ingress support for clusters displayed on main and “Switch Cluster” UIs
  • Consistent Health Score throughout the user experience
  • Fix for drill down issue on certain label names
  • Restoration of “Tenancy costs” chip on Allocation page shared resources
  • Improved readability for long names in Allocation and Asset pages
  • Better status for empty spot recommendations
  • Grafana dashboard updates to address a number of data issues - #1094

Performance

  • Improved performance when paging results on Athena queries
  • Faster load times in allocations that load many external assets - kubecost/cost-model#960

Stability/Testing

  • Add store PVC default to avoid ephemeral storage issues - #1121
  • Fixed nil panic bug when running on K3s clusters on GCP - kubecost/cost-model#957
Compare
Choose a tag to compare
  • This release containers a workaround for kubecost/cost-model#964 where certain root disk sizes would be incorrectly emitted by node-exporter. All ongoing metrics emission will be handled by this fix, but if you were impacted and are seeing extremely high root disk sizes/costs, you will want to run a rebuild of our asset ETL cache with
    /model/etl/asset/repair?window=7d
Compare
Choose a tag to compare
  • Fix issue where some SAML setups would filter all values.
  • Fix issue where rare k8s API issues would cause very large prices.
Compare
Choose a tag to compare
  • Fix issue where assets page would be blank for an unset currency code.