v1.99.0-rc.8

v1.99.0-rc.8 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.99 is officially released.

After taking v1.99 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.99 release. The estimated release date for v1.99 is Tuesday, December 13.

v1.99.0-rc.0

v1.99.0-rc.0 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.99 is officially released.

After taking v1.99 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.99 release. The estimated release date for v1.99 is Tuesday, December 13th.

v1.98.0

v1.98.0 Release Notes

Deprecations

• PodSecurityPolicies: Opencost and Kubecost now both support kubernetes v1.25. However, by default, certain PodSecurityPolicies are still installed. These will be removed in v1.99 to streamline installation on v1.25. Meanwhile, you can still install Kubecost on v1.25
  via this helm chart with the following instructions: #1773 (comment)
• We will be deprecating older versions of Kubecost. Our earliest supported version is v1.70.0

Major

• Launched Kubecost Cloud beta
• Launched free unlimited clusters
• Launch new UI Beta
• *Doc Container Request Right Sizing recommendation API V2 Beta
• Persistent Volume Right Sizing insights beta
• ETL Query Scaling APIs to improve allocation filtering performance at scale!
• 1700 Launch of federated ETL pipeline Beta Docs
• Reduced memory consumption in cascading allocation ETL

Note: * Percentile-based request right-sizing recommendations are available in the Beta UI, but only if the "ContainerStats" data pipeline is enabled. This pipeline may have trouble at high data scales while it is in the early stages of development -- please carefully review the relevant documentation: https://github.com/kubecost/docs/blob/main/api-request-right-sizing-v2.md

Minor

• Improved initial UI responsiveness by reducing overall UI bundle size by 2%
• 366 Add support for Workload Identities for GCP Big Query Config
• Add backend caching and single flighting to the ETL assets endpoint
• Re introduced the network.html page
• Created an EventLog representation of the ETL events to avoid storing the Data field longer than expected to reduce memory usage
• Invoke shallow clones in aggregate store to save memory
• Added an optional kubecostModel.softMemoryLimit field which sets the GOMEMLIMIT flag for the cost-model backend

Bugs

• 1605 Improved OIDC user authentication.
• 1694 Fixed Read-only users being able to access the diagnostics page and perform operations
• 1730 Fixed bug in abandoned workloads
• 15 Fixed an issue where the context switcher would strip https and always use http for new contexts
• 1760 Stops logging invalid sharing coefficient when coeff equals zero
• 1074 Fixed bug preventing adding multiple filters of the same type on Allocatio
• 1675 Fixed duplicate and unexpected node chips in k8s Assets view
• Updated golang.org/x/net and /x/text to fix CVEs
• Removed bad property from button that caused etl-status page to crash
• Fixed a bug with the S3 select integration where multiple csv per month caused line items in the separate CSV file to not be ingested
• Fixed bugs with summary allocation and CSV formats where accumulation was mutating cache
• Fixed GetETLCloudRunWindowDays function checking the wrong ENV
• Fixed broken Azure Service Tagging
• Fix error in Bug Report's Thanos queries
• Fixed bug causing shared labels/namespaces/overhead persist when opening a report from reports
• Fix panic in accumulateAttributes
• Added support for unpartitioned GCP account labels
• Fixed duplicate names causing issues with graphs
• Update hard coded references to product tiers to use useProductTier
• Fixed bug causing custom labels filters to not work as expected on details page

Community Contributions

• 1741 Updates the loopback address in the in-zone classification to include the entire loopback range as registered in IANA

OpenCost

• 1332 Fixed v1.95.0 broken URLs
• 1461 Disabled window filters which caused circular deps
• 1460 Added ability to filter by annotations for owner, department, team, product, and environment
• 1457 Added ability to filter by annotations for owner, department, team, product, and environment
• 1454 Fixed a bug which caused Azure provider to ignore Azure Storage configurations from secret
• 1452 Added support for K8s v1.25+ by removing dependency on autoscaling/v2beta1 API
• 1450 Fixed an issue where updating configuration (e.g. currency) on Azure would deadlock
• 1443 Misc. changes to OpenCost concepts in order to support Federated ETL
• 1438 Made usage of Disk resilient to absence of metrics
• 1432 Added methods for Cloud Cost types, removed name and times from cloud cost types
• 1427 Support MSI authentication for Azure by always returning config values if secret not found
• 1425 Added test to ensure AllocationSet.accumulate is equivalent to AllocationSetRange.Accumulate
• 1421 Added GP3 volumes support in AWS
• 1418 Updated the go version to 1.19 and removed the pkg/util/atomic primitives in favor of stdlib atomic primitives introduced in go 1.19
• 1417 Allows users of the UI to override the path to use when connecting to the cost model
• 1415 Adds a Window getter to AssetSet and AllocationSet for compatibility with generic constraints
• 1404 Adds a NewAccumulation() method on the SetRange ETL data structures which provides an Accumulate method which does not mutate.
• 1403 Pulls in PV usage metrics into ETL
• 1401 parsewindow functions returns window only in UTC when window requests are time stamped and developer is testing functions locally with KCM
• 1400 Improves usage of v1 filter function
• 1397 Introduces a new Prometheus context name for the new container stats pipeline (closed source)
• 1381 Exposes functions needed to calculate efficiencies over summaryAllocationSets

v1.98.0-rc.4

v1.98.0-rc.4 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.98 is officially released.

After taking v1.98 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.98 release. The estimated release date for v1.98 is Wednesday, November 2nd.

v1.98.0-rc.3

v1.98.0-rc.3 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.98 is officially released.

After taking v1.98 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.98 release. The estimated release date for v1.98 is Monday, November 2nd.

v1.98.0-rc.2

v1.98.0-rc.2 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.98 is officially released.

After taking v1.98 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.98 release. The estimated release date for v1.98 is Monday, November 2nd.

v1.98.0-rc.1

v1.98.0-rc.1 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.98 is officially released.

After taking v1.98 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.98 release. The estimated release date for v1.98 is Monday, November 2nd.

v1.97.0

v1.97.0 Release Notes

Major

• 324 Request right-sizing recommendation API V2
• Launched alpha version of the new Overview homepage available at /overview-new

Minor

• 1491 Added Helm template value to disable IPv6
• 1679 Added new values file for AWS AMP integration
• 1652 Added the ability to add affinity value to prometheus/node-exporter
• Added function to prevent a mixture of chipsets from being suggested in the complex recommendations section of cluster sizing
• Added the ability to copy Orphaned Resources list to the clipboard
• Improved accuracy for cloud usage data on AWS by filtering based on usage rather than the presence of resource ID
• Added the ability to drill down when aggregating by label on the Cost Allocation page
• Cleaned up the report builder UI in Advanced Reports
• Added microcopy for empty tables and states to improve the user experience when data is not available

Bugs

• 1598 Updated error shown on the cluster-sizing recommendation page
• 1644 Fixed annotations not showing up in aggregations. This fix allows users to filter by annotations as well as labels. Works for owner, product, team, department, environment.
• 1633 Bumps the Prometheus scrape timeout to 60 seconds to avoid issues where, in high-scale clusters, Kubecost can take additional time to respond to Prometheus.
• 1639 Fixed issue where graphs would not properly display while using Grafana behind an ingress
• Fixed a bug whereUnutilized Disk Savings page could fail to load for multi-cluster environments
• Limit the read to the first value of the nf_conntrack_acct to ensure accurate comparisons
• Fixed an issue which caused new aggregations not to save within Advanced Reports
• Fixed a bug that prevented updating GCP cloud provider integrations from the UI
• Fixed a bug where hours run in Assets was displaying in minutes instead of hours
• Added dynamic namespace for bug report generation
• Updated broken links on Cost Allocation page
• Added protection against nil panics when accessing cloud usage data and provider info
• Added a nil check to Cloud ETL Store, given the Provider API doesn't guarantee a non-nil SetRange on QueryCloudUsage
• Fixed typos on 1.96.0

Community Contributions

• 1643 Added support for Amazon managed service for Prometheus

OpenCost

• 1392 Sanitized label name when filtering on alias
• 1388 Typo cost-model/ui -> opencost/ui
• 1376 Allowed full matches on cloud type assets
• 1373 Fixed bad namespace on Prometheus environment variable
• 1372 Added real world example to unit tests
• 1367 String Bank Performance Gains
• 1364 Fixed typo
• 1363 Update PV and LB to use Unmounted
• 1361 Fix PV alloc recon, alloc PV marshaling
• 1359 Implement .Equals() for the AllocationFilter interface

v1.97.0-rc.5

v1.97.0-rc.5 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.97 is officially released.

After taking v1.97 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.97 release. We are currently at our yearly company retreat, and the estimated release date for v1.97 is Monday, September 19th.

v1.97.0-rc.4

v1.97.0-rc.4 is a pre-release.

It is to help gather feedback from the community and give users a chance to test Kubecost before v1.97 is officially released.

After taking v1.97 for a test drive, if you have any feedback, it would be great if you could open an issue here.

The official change log will come out with the v1.97 release. We are currently at our yearly company retreat, and the estimated release date for v1.97 is Monday, September 19th.