github kubecost/cost-analyzer-helm-chart v1.91.0

latest releases: v2.5.0-rc.1, v2.5.0-rc.0, v2.4.2...
2 years ago

Configuration/Architecture

  • Fix: ETL configuration was not properly passing remote bucket configuration
  • Support for disabling emission of specific metrics by Kubecost
  • Grafana dashboards cluster-util and label-cost-util: Fixed N/A values and errors
  • Vulnerability "CVE-2021-3538" in dependency "github.com/satori/go.uuid" fixed
  • The open source container is now interchangeable with the closed source container
  • Helm chart supports custom args for cost-model and network-costs
  • Kubecost now has a custom 504 error page to help debug 504 error cases where multiple gateway/proxy layers are present
  • Configurable SAML session timeout and logout redirect URL
  • Misc dependency version bumps
  • Add IPv6 support
  • Network Costs "updateStrategy" can be set via Helm values
  • The KSM cluster role binding will not be deployed if KSM is disabled via "prometheus.kube-state-metrics.disabled=true"

Cost

  • Fix: PV costs are no longer incorrectly duplicated in the event of a PVC being shared by multiple pods
  • Fix: Savings summary and request sizing no longer double-apply discounts
  • Fix: Idle costs were incorrect when filters are applied and adjustments are present
  • Fix: Prometheus and Thanos clients handle rate limit responses, including AWS AMP 400 special-case throttling
  • Fix: Avoid a nil panic in an Athena query failure edge case
  • The abandoned workloads endpoint will fail more quickly in the event of bad data
  • Reduced memory impact on startup for cloud asset/reconciliation processes
  • Reduced memory impact for AWS users with cloud asset/reconciliation enabled
  • Support for alert testing
  • Fix: "/model/explainCA" no longer returns garbled error messages
  • Fix: "/model/explainCA" returns HTTP error statuses for errors
  • Kubecost's backing schema can be upgraded without rebuilding the ETL
  • Fix: recurring alert duration now reported correctly
  • Allocation API supports the "accumulateBy" parameter

Frontend

  • Allocation: New treemap visualization of allocation data
  • Diagnostics: Max and average node count, as calculated by Kubecost, visible
  • Inspect: Support for label aggregation
  • Allocation: Page now includes a link to Kubecost official documentation
  • Savings: Fixed some mangled URLs
  • Savings: savings.html should load more quickly and more stably
  • Alerts: UI reworked for usability and adds alert testing functionality
  • SAML: Logout button
  • Overview + Savings: savings estimate being slow will not break the page (instead loading partially)
  • Allocation: "NaN%" is replaced with "-" for external_resources
  • Inspect: Useful failure message when data isn't available to display
  • Allocation: On the details drilldown for an individual pod, the row values were being too-aggressively rounded to 2 decimal places while displaying 3. The values are now rounded to 3 decimal places.
  • Cloud connection status: fixed "No Connection" status bug
  • Namespace: disable unsupported modals (update namespace and budget) with tooltip explanation

Bugs fixed in this release:

Don't miss a new cost-analyzer-helm-chart release

NewReleases is sending notifications on new releases.