Feature Overview (aka. Goal Summary)

Intelligent Node Selection for Must-Gather Pods
Enhance oc adm must-gather to avoid scheduling pods on NotReady or unreachable nodes. Currently, in disconnected or degraded environments, must-gather pods often get stuck in Pending state when scheduled on tainted or unavailable nodes. This enhancement ensures must-gather pods are reliably scheduled on healthy, Ready nodes—improving diagnostic reliability and reducing manual intervention.

Goals (aka. Expected User Outcomes)

• 🛑 Prevent Diagnostic Failures: Must-gather avoids NotReady or NoExecute-tainted nodes during pod scheduling to prevent stuck or incomplete data collection.

• ⚙️ Reliable Automation in Disconnected Setups: Users in air-gapped or MNO environments can run must-gather without manual --node-name intervention.

• 🔍 Accurate and Complete Cluster Insights: Ensures must-gather consistently collects logs and cluster state even during partial outages.

• 🧠 Health-Aware Pod Scheduling: Must-gather automatically selects a healthy Ready control-plane or fallback node.

• 🔁 Cross-Version Consistency: Delivers uniform behavior across OpenShift 4.14.z, 4.16.z, and 4.18+ to cover common production deployments.