v1.51.2

• Display non-local deployments on namespace page in multi-cluster view
• Show unallocated resources in Allocation page
• Expose cluster sizing feature to more users
• Fix open Azure GPU issue to address label change
• Fix SAML redirect bug for users without product access
• Better bug reporting capabilities
• Add more audit metrics
• Improved profiling logs

v1.51.1

• Fix L1 caching edge case with shared costs
• Turndown improvements
• Minor frontend performance improvements

v1.51.0

• Bring new smart turndown features to more users
• Expose cluster right-size functionality to feature-council group
• New access path for real-only SAML users
• Support AttributeName vectors in SAML data
• Caching optimizations for namespace data
• Add pv storageclass option
• Add Allocation totals
• Various bug fixes, including NL Azure node prices, network allocation time window/durations

v1.50.4

• GPU fix for potential label duplication scenario
• Efs volume fix
• More lock safety for race condition on price download
• SAML cookie deletion/expiration issue
• Adding shared labels configuration to settings
• Move default shared cost allocation to proportionate from fixed -- i.e. using sharedSplit=weighted, https://github.com/kubecost/docs/blob/master/allocation-api.md#aggregated-cost-model-api

v1.50.3

• Performance improvements on certain clusters, e.g. missing nodes
• New profiling logs for performance & resource consumption
• Deep linking on Allocation view
• Race condition fix on product configs provided at install time
• New audit tool functionality
• New bug report feature
• Additional configuration for certain SAML providers
• Other various bug fixes

v1.50.2

• Move additional product configurations to install time, i.e. values.yaml
• Expose Statefulset costs
• New abandoned workloads + orphaned resources views
• Add new audit tools
• Better logs for unexpected model results, e.g. NaNs
• More granular cost model timer logs
• Fixes to several edge case issues
• Address network issue with different conntrack configurations

v1.49.3

• Introduce initial statefulset metrics
• Expose additional alert configuration options
• Add ksm version check diagnostic
• Provide workaround for Thanos memory management issue -- thanos-io/thanos#1471
• Fix various other minor bugs

v1.49.2

• Bust new frontend cache for config changes
• Drop additional unused metrics
• Fix time series unsafe access bug

v1.49.0

• Expose cost model label filters
• Optimize default Thanos + Prometheus memory usage during metric caching
• Upgrade Thanos to v0.8.1
• Add additional reserved instance calculations
• Fix n2 calculation bug
• Add new pricing data refresh logic
• Create list of default Prometheus metrics to drop

v1.48.1

• Fix efficiency calculation issue: opencost/opencost#238
• Improve cold start experience
• Add email alert debug mode + logs