v2.8.1 - 17/Mar/2026
Bug Fixes
-
Resource Override Precedence: Resolved an issue introduced in version 2.8.0 where the sizing profile incorrectly took precedence over user-specified resource overrides. User-defined settings now correctly override defaults.
-
License Validation via External Secrets: Fixed a regression from version 2.8.0 that caused license validation to fail when credentials were retrieved using the
fromExternalSecretconfiguration. -
Actor System Corruption: Fixed an intermittent "Actor System Corrupted" error that previously required a manual pod restart. As part of this fix, the Java Security Manager has been disabled, which also eliminates associated deprecation warnings.
-
Helm "Global" Mode Configuration: Fixed a bug introduced in 2.3.0/2.8.0 where user-provided overrides for resources, volume sizes, replicas, and retention were not always respected in the new global configuration mode. Note: Legacy profile-based configurations were not affected by this issue.
-
Admin Password Updates: Resolved an issue where changes to the admin password in Helm values sometimes required a manual restart of the server or API pods to take effect.
-
Backup CLI (sts-backup) Improvements:
-
Port-Forwarding: The tool now utilizes dynamically allocated local ports for Kubernetes port-forwarding to avoid port conflicts.
-
Restore Reliability: Removed the timeout for restore operations to ensure a smoother user experience during large StackGraph restorations.
-