v2.11.0 - 14/Sep/2026

Release Notes: SUSE® Observability Helm Chart v2.11.0

Component Versions

Component Version

SUSE Observability

2.11.0

SUSE Observability Agent

1.6.1

SUSE Observability CLI

3.9.0

SUSE Observability - StackPacks 2.0

Overview

StackPacks 2.0 makes it possible to integrate any technology into SUSE Observability through configuration alone. A StackPack maps OpenTelemetry traces, metrics, and logs into topology, and extends the UI with overview pages, detail-page widgets, menu items, and dashboards. StackPacks are published and shared through an open, community-driven Git marketplace and are automatically exposed to AI assistants via the SUSE Observability MCP Server.

What’s New

  • Configuration-Driven Ingestion: Map raw OpenTelemetry traces, metrics, and logs into topology maps and dashboards through configuration only. No platform code changes or custom plugins are required, so a new technology can be onboarded in minutes rather than months.

  • Extensible UI: StackPacks can now add and update overview pages, widgets on detail pages, menu items, and dashboards. This lets a StackPack deliver a complete, technology-specific experience rather than topology alone.

  • Community-Driven StackPack Marketplace: StackPacks can be built, extended, and shared through a public GitHub repository. StackPacks are based on open, vendor-neutral schema standards, so contributions from the community and partners work across environments.

Example Use Case

Engineering teams deploying AI frameworks such as vLLM, Ollama, or LangChain on Kubernetes can ship a StackPack that maps their OTel data into topology, adds dedicated dashboards and detail-page widgets, and exposes the resulting metrics to AI assistants for root-cause analysis.

Getting Started

  • Browse and install StackPacks from the public GitHub marketplace.

  • Author or extend a StackPack using the open schema to map your OTel data and define UI extensions. (Fun fact: StackPacks can be fully authored with current leading-edge AI models.)

  • Create a pull request to share your StackPacks with the community.

Resources

New Features & Enhancements

  • Stackpacks: Added the ability to install specific versions of a Stackpack rather than defaulting to the latest version.

  • Tracing: Added detailed inspection for trace spans and span events.

  • Kubernetes & Helm: Trust stores can now be loaded from user-managed Kubernetes secrets, preventing Helm release secrets from exceeding the 1MB limit during upgrades.

  • Agent Configuration: podAnnotations and podLabels can now be set per agent workload (clusterAgent, nodeAgent, checksAgent, logsAgent, and remoteKubeCache), allowing monitors to be disabled on specific agent pods via Helm values.

  • Integrations: Configuration items shipped by integrations now have immutable ownership and cannot be locked or unlocked. This applies to monitors, dashboards, views, main menu groups, and other configuration items.

  • StackGraph Backup & Restore: Added a new stackgraph-v2 backup and restore solution with incremental backups and improved performance.

  • Topology Perspective: Reduced CPU usage when using the "Find component" feature.

  • Agent Performance: Reduced the CPU limit for the logs agent.

  • Backup Performance: Limited S3Proxy to 32 threads to avoid out-of-memory errors when multiple restore operations run in parallel.

  • Storage Configuration: Per-volume storageClass values now override global.storageClass for Kafka, ZooKeeper, ClickHouse, and Elasticsearch volumes.

Bug Fixes

  • Platform & CLI: Fixed cloning issues for PromQL-based monitors in the CLI.

  • Topology Snapshot: Fixed an issue where the topology sync initial snapshot prevented switching the active host.

  • Monitors: Fixed an issue where the Out of Memory monitor disappeared due to an incorrect label value.

  • Security & Compliance: Updated all Helm chart containers to run with the securityContext required by the restricted Pod Security Standard, preventing admission failures on CIS-hardened RKE2 or Rancher-restricted clusters.

  • Elasticsearch: Fixed startup failures on FIPS-enabled nodes by including the complete NSS runtime and a corrected PKCS#11 FIPS configuration.

  • Backup Infrastructure:

    • Set concurrencyPolicy: Forbid on the backup-init CronJob.

    • Updated the backup-init Job and CronJob to use restartPolicy: OnFailure, so failed attempts no longer leave orphan Pods behind.

    • Hourly VictoriaMetrics backups now refuse to overwrite a completed backup when the local instance has no data older than 24 hours. This prevents a from-scratch redeploy from destroying its backup before data is restored. Configure the threshold per instance with victoria-metrics-<instance>.backup.minHistorySeconds (0 disables the check). Refused backups exit non-zero, allowing the existing VictoriaMetricsBackupsPerformed alert to report the failure within about three hours.

    • The Backup CLI now retries when port-forwarding fails.

    • The Backup CLI now checks Elasticsearch restore operation status correctly and waits until all indices and shards are healthy.

  • Topology Perspective: Fixed component focus when using the "Find component" feature.

  • User Interface: Prevented drawer forms from automatically restoring after accidental closure with the Escape key or overlay clicks.

Agent Bug Fixes

  • Agent Configuration Rendering: Fixed invalid YAML generation when setting global.extraLabels or global.extraAnnotations, which previously caused STS_CLUSTER_NAME and the receiver URL to be dropped silently.