v1.66.0

App:

• Use new allocation API from checks/alerts pod
• Savings performance improvements
• Multicluster performance improvements for Overview
• Monthly rate estimation improvements
• Reduce cadvisor metrics collected with whitelist -- #608
• Use appropriate k8s apis for Daemonset,PriorityClass, NetworkPolicy, and PodSecurityPolicy based on k8s version -- #591
• Disable PriorityClass by default to avoid helm bug -- #616
• Support option to disable network-cost traffic logs -- #609
• Remove unnecessary cost-model logs

v1.65.0

App:

• Move savings to to multi-cluster view on Enterprise tier
• Leverage new recording rules added in v1.64
• Migrate checks image to quay image repo (both repos still supported, but gcr is planned to be deprecated)
• Upgrade checks pod base image
• Move ksm and node-exporter to metric whitelist - #602
• Upgrade Grafana -- #600
• Add additional frontend sanitization for config values
• Migrate Overview to show allocation by Controller & Service instead of Deployment and Product
• Directly support GCP foreign currency exchange
• Split allocation over multiple pods attaching to a PVC
• Emit cluster info metrics -- opencost/opencost#530
• Support aggregating data by multiple labels simultaneously on the cost allocation API — #598
• Fix: idle cost being filtered incorrectly on ETL pipeline
• Fix: network cost labeling issue for certain cross-region traffic
• Fix: properly support OOC from external masterpayer account -- opencost/opencost#532:
• Support ignoring Prometheus cert — values

v1.64.0

App

• New assets pipeline -- available at /assets.html and accessible from Overview page
• Initial release of cloud bill reconciliation -- support matching node costs from AWS
• Support node-class based sustained use discounts in GCP
• Add Cluster Management Fees charged by EKS and GKE to new Assets pipeline
• Improved SAML debugging statements
• Fix SAML redirect loop when multiple SAML protected clusters exist and one isn't authenticated
• Fix email copy where some alerts weren't properly displaying the correct linkback url
• Support special characters and upper case values in external cloud tags

Helm Chart

• Add new recording rules for multi-cluster savings --https://github.com/kubecost/cost-analyzer-helm-chart/pull/587/files
• Multi-cluster basic auth fix to support empty secret -- https://github.com/kubecost/cost-analyzer-helm-chart/pull/588/files

v1.63.1

App

• Fix: update copy when certain AWS key failures displayed back to user
• Fix: aggregate metrics for Cronjob allocations
• Fix: Handle special label characters in ETL pipeline
• Fix: Prevent orphaned resource detection errors from blocking page loads
• Feature: Label filters: set <labelname>= __unallocated__ to filter by all allocations where <labelname> is not set

Helm Chart

• Fix: Resolve issues with AWS service key secret creation

v1.63.0

Helm Chart:

• Increase default query.max-samples for Prometheus #570
• Pass cluster context at install-time -- Impact default resource recommendations for request-sizing, etc #568
• Add auth support for durable storage endpoint #567
• Support different release names on checks pod

Application:

• Add new assets view -- Available at /assets.html
• Add AWS key validation -- new Settings card to view AWS key permissions
• Fix issue where memory requests recommendation could be skewed by many pod restarts
• Fix issue where spot nodes leaving the fleet might continue to be billed
• Use local version of jquery UI for Allocation view
• Add helm values to bug report
• Remove case sensitivity requirement in custom CSV asset pipeline
• Update AWS Savings Plan implementation
• Add pod prefix filter
• Add support for cost aggregation by jobs

v1.62.0

App:

• Support timezones on ETL pipeline
• Fix panic on empty time parameter string
• Fix issue where new node type prices in GKE were not picked up until pod restart
• Reduce duplicate logs
• Extend default ETL pipeline build times to 120 days
  Helm Chart:
• Add support for service key secret creation via helm
• Add values.yaml to debug report
• Support PodMonitor on network costs pod
• Amazon Web Store fixes
• Add resources block to init container

v1.61.3

App:

• Fix issue with default values.yaml not correctly adding datasource to grafana
  -- NOTE: A restart of the grafana pod and a hard refresh of the grafana UI may be required to pick up these changes.
• Fix issue with custom prices less than 0.01/hour not supported in values.yaml
• Allow integers to be supplied in values.yaml prices
• Support throttling of prometheus queries in application
• Improve accuracy of cluster costs on startup
• Add Canadian dollar as currency

v1.61.2

App:

• Fix frontend bug on namespace.html
• Minor change to "today" rate projection calculation
• Add support on the CSV provider for disambiguation of nodes with identical provider IDs by region.

v1.61.1

App:

• Fix rate interpolation issue
• Fix startup javascript error for new Reports page

v1.61.0

Helm:

• Fix issue with master payor ARN
• Update default Grafana data source

App:

• Move saved reports to GA!
• Make cluster metrics on Overview show costs from multi-cluster
• Refresh spot data feed more frequently
• Use ETL pipeline on custom windows
• Interpolate hourly time series when no data collected
• Properly handle delayed ETL data on multi-cluster