V2.2.3
Fixed
- Fix issues with creating Cloud provider integrations via the UI.
- Fix an issue with continuous right-sizing calls not passing seconds in timestamp to API.
- Fix an issue with Abandoned Workloads where selecting the date range would override the threshold.
- Fix an issue where the cluster inspect page would incorrectly show used/requested cpu cores as millicores.
- Fix issues where the overview network tables did not match the data shown on click-through.
- Fix an issue with AWS workload identities.
- Fix network insights API returning an empty response in a single cluster install.
- Fix an issue creating scheduled reports via the helm config map resulting in next run time not being set.
- Fix an error when aggregating by label and annotation when idle costs are not hidden.
- Fix an error when querying assets with empty asset sets giving boundary errors.
- Fixes idle sharing which was broken for CPU, GPU, and RAM on the Allocation API
- Fixes a bug where idle GPUs were not properly accounted
- Fix an error on asset ingestion allowing spikes in pv costs.
Helm Changes
#3324 Disable helm-rollout-restart if cicd=true