Skip to content

Burst: Scale Out / In

Purpose: Temporarily add/remove cloud capacity based on load or scheduled peaks. Owner: Platform SRE · When: Alert threshold crossed or scheduled peak.

Pre‑requisites

  • Authorization to burst (policy/Decision Service/schedule).
  • Networking and cluster access in place.

Steps

1) Scale out (add capacity) 

export CLOUD=gcp   # or azure
make burst.scale.up CLOUD=$CLOUD   | tee "output/logs/terraform/$(date -Iseconds)_burst_up_${CLOUD}.log"

2) Validate workload placement 

make burst.validate
kubectl top nodes   | tee "output/artifacts/showcases/kubernetes-autoscaling/$(date -Iseconds)_top_nodes_${CLOUD}.txt"
kubectl get pods -A -o wide   | tee "output/artifacts/showcases/kubernetes-autoscaling/$(date -Iseconds)_pods_${CLOUD}.txt"

3) Scale in (remove surge) 

make burst.scale.down CLOUD=$CLOUD   | tee "output/logs/terraform/$(date -Iseconds)_burst_down_${CLOUD}.log"

Evidence

  • HPA/VPA events & graphs → output/artifacts/observability/images/.
  • Terraform logs → output/logs/terraform/.