Q: Explain how the Kubernetes scheduler makes a placement decision for a new pod.
The scheduler goes through two phases:
#Kubernetes #Advanced Scenarios #L3 #Container Orchestration #K8s
🎙️ Candidate Opening & Architectural Context
""In one of our high-traffic production clusters, our SRE team handled this incident using a standardized runbook. When addressing this question, I walk the interviewer through our production incident runbook: isolating the blast radius, checking diagnostic logs and metrics, and applying a safe fix.""
Advertisement
🛠️ Production Runbook & Step-by-Step Resolution
1️⃣
Initial Diagnostics & Root Cause Analysis
The scheduler goes through two phases:
- Does the node have enough CPU and memory?
- Does the node match the pod's
nodeSelector? - Does the pod tolerate the node's taints?
- Are the required volumes available in that zone?
- Does the pod's affinity/anti-affinity rule allow this node?
2️⃣
Remediation & Permanent Safeguards
Phase 1: Filtering (Predicates) Eliminate nodes that can't run the pod: Phase 2: Scoring (Priorities) Rank the remaining nodes: The highest-scoring node wins. If tied, a random one is picked. The scheduler then writes the chosen node name into the pod's spec (nodeName field), and the kubelet on that node picks it up and starts the pod.
- Least requested resources (prefer emptier nodes for better bin-packing).
- Affinity preference scores.
- Topology spread.
- Image locality (prefer nodes that already have the image pulled).
💡 The Senior SRE Gold Nugget (Key Architectural Takeaway)
"Pro-Tip: Does the node have enough CPU and memory?."
⚡ 60-Second Elevator Pitch Talking Points
- Does the node have enough CPU and memory?
- Does the node match the pod's nodeSelector?
- Does the pod tolerate the node's taints?
Advertisement