#2 Target Porter application nodes by default

Summary

• schedule the node collector only on Porter Application nodes by default
• keep Kubernetes events and workload-state collection cluster-wide
• support an explicit all-node override through Helm values
• document coexistence with Porter built-in monitoring and the pod-capacity tradeoff
• release chart version 0.1.1

Testing

• bash scripts/validate.sh
• shellcheck scripts/*.sh charts/superlog-otel/tests/*.sh
• actionlint
• git diff --check

---

Summary by cubic

Default the superlog-otel node agent to run only on Porter Application nodes to avoid pod pressure on System nodes, while keeping cluster-wide events and workload metrics. Adds an all-nodes opt-in via Helm values and bumps the chart to 0.1.1.

• New Features
  • DaemonSet now uses node affinity on porter.run/workload-kind: application by default.
  • New value collectors.agent.porterNodeGroups (default ["application"]); set to [] to target all Linux nodes.
  • Updated docs on Porter node groups, scheduling tradeoffs, and coexistence with built-in monitoring.
  • Tests updated to assert defaults, override behavior, and chart version.