FFORGE//RS
← Roadmap

distributed systems / LEVEL 3

The rollout that dropped below capacity

EST.40 MIN
01

THEORY / RETRIEVAL

What to restore

  • Calculate upper pod count and lower ready capacity
  • Reject rollout budgets that can remove every replica
  • Separate readiness from process liveness

Rollout settings are capacity constraints

For an integer exercise, maxSurge raises the maximum total pods during rollout while maxUnavailable lowers the minimum ready replicas. The safe values depend on cluster headroom, per-pod throughput, disruption budgets and whether readiness accurately represents ability to serve.

Readiness controls traffic; liveness controls restart

A failed readiness probe removes a pod from service endpoints without necessarily restarting it. A liveness failure restarts the container. Reusing an aggressive dependency check for liveness can turn a downstream incident into a restart storm.

CHECKPOINT

A pod has started but has not loaded routing state. Which probe should keep it out of Service traffic?

ISOLATED RUST 1.96
src/lib.rsEDIT

02 / CODE REVIEW

Make the code safe

Assess an integer rolling-update budget. replicas and per_pod_rps must be positive, max_unavailable must be less than replicas, and all arithmetic is checked. Report max total pods, minimum ready pods/capacity, and whether that lower bound serves target_rps.

Initializing editor…
CLOUD SANDBOXnetwork off · 256 MB · 12 s
1 / 64 KB
OUTPUT
Runner is waiting for a submission.