v2.6.3 - 25/Nov/2025

HDFS Upgrade: HDFS (Hadoop Distributed File System) and its associated dependencies have been upgraded.

StackPack: Partial Topology Sync Monitor: A new monitor has been added to the StackState StackPack to alert on partial Topology Synchronization snapshots.

vmagent Resource Increase: The memory and CPU resource requirements for the vmagent component have been increased in the 4000-ha profile.

Image Upgrades:

OpenTelemetry Metric Scoping: Fixed a critical issue where metrics ingested via the OpenTelemetry collector were missing the scope label. This prevented scoped users from being able to observe these metrics.

Metric Explorer Sorting: The Metric Explorer now uses numerical sorting for values in the value column.

Platform: StackGraph Corruption (Timed-Out Transactions): Fixed a StackGraph corruption issue where data from timed-out transactions that should have been rolled back could inadvertently reappear.

Platform: State Pod Validation: Added additional data validation and logging to the state pod for improved stability and debugging.

StackGraph: Edge Deletion Invariant: Added an invariant to prevent inconsistent edge references when performing a delete edge operation in StackGraph.

StackGraph Integrity Verifier: An experimental perpetual integrity verifier has been added for StackGraph. It can be enabled by setting hbase.console.integrity.enabled=true.

StackPack Remediation Guides: Fixed several remediation guides within the SUSE Observability stackpack that incorrectly referenced tags instead of the correct term, labels.

Duplicate OpenTelemetry StackPack: Removed a duplicate OpenTelemetry stackpack installation.

Platform: Agent Restart Snapshot Loop: Fixed an issue where a restart of an agent could cause the 'active snapshot' to continuously occur.

Platform: Kafka JMX OOM Fix: Resolved an Out-Of-Memory (OOM) issue for the Kafka JMX container on RKE2 Kubernetes versions 1.31 and 1.30.