Add kubernetes-deployment: probes, shutdown, JVM ergonomics, HPA

Companion code for "Deploying Spring Boot 4 on Kubernetes: Probes, Graceful Shutdown, Limits and JVM Ergonomics". A dependency outage under three probe-group setups, a rolling restart under load four ways (three runs each), the JVM's ergonomic choices for nine pod shapes, one GC-heavy load under five CPU limits with throttling counters, and an HPA driven by a Micrometer gauge through prometheus-adapter. Measured on k3s v1.36.4. Co-Authored-By: Claude Opus 5 <[email protected]> Claude-Session: https://claude.ai/code/session_01C3TETMrqVUWeFkNtz3Jbo3
2026-09-11 17:12:10 +00:00
parent 644da9e65e
commit a065696478
72 changed files with 28415 additions and 2 deletions
@@ -0,0 +1,57 @@
+#!/usr/bin/env bash
+# HorizontalPodAutoscaler on a Micrometer gauge through Prometheus + prometheus-adapter.
+# Load: 2 clients, then 30 from t=20 s to t=140 s, then none; each request holds for 500 ms.
+# -> docs/output/hpa-custom-metric.txt, hpa-custom-metrics-api.txt
+set -uo pipefail
+source "$(dirname "$0")/env.sh"
+H="$MODULE_DIR/k8s/hpa"
+kubectl apply -f "$H/prometheus.yaml" -f "$H/prometheus-adapter.yaml" > /dev/null
+kubectl -n monitoring rollout status deploy/prometheus --timeout=120s > /dev/null
+kubectl -n monitoring rollout status deploy/prometheus-adapter --timeout=120s > /dev/null
+kubectl -n "$NS" scale deploy/orders --replicas=1 > /dev/null
+kubectl -n "$NS" rollout status deploy/orders --timeout=120s > /dev/null
+API=/apis/custom.metrics.k8s.io/v1beta1/namespaces/demo/pods/%2A/app_inflight_requests
+for _ in $(seq 1 40); do kubectl get --raw "$API" 2>/dev/null | grep -q '"items":\[{' && break; sleep 5; done
+{
+  echo "# What Micrometer exports (one pod):"
+  echo "\$ curl <pod>:8080/actuator/prometheus | grep app_inflight"
+  curl -s "$(kubectl -n "$NS" get pod -l app=orders -o jsonpath='{.items[0].status.podIP}'):8080/actuator/prometheus" | grep app_inflight
+  echo
+  echo "# What the HPA controller sees through the custom metrics API:"
+  echo "\$ kubectl get --raw $API"
+  kubectl get --raw "$API" | python3 -m json.tool
+} > "$OUT/hpa-custom-metrics-api.txt"
+
+kubectl apply -f "$H/hpa.yaml" > /dev/null
+sleep 20
+kubectl -n "$NS" delete events --field-selector involvedObject.kind=HorizontalPodAutoscaler > /dev/null 2>&1
+"$MODULE_DIR/scripts/loadgen.sh" lg-hpa "http://orders:8080/work?ms=500" "0:2,20:30,140:0" 230
+kubectl -n "$NS" wait --for=condition=Ready pod/lg-hpa --timeout=60s > /dev/null
+start=$(date +%s)
+{
+  echo "# HPA orders: target app_inflight_requests averageValue 5, min 1, max 4; scaleDown stabilization 30 s"
+  echo "# load: t=0 2 clients, t=20 30 clients, t=140 0 clients; GET /work?ms=500"
+  echo
+  printf '%-6s %-10s %-9s %-8s %s\n' "t" "clients" "metric" "desired" "ready pods"
+  while [ $(( $(date +%s) - start )) -lt 235 ]; do
+    t=$(( $(date +%s) - start ))
+    clients=2; [ $t -ge 20 ] && clients=30; [ $t -ge 140 ] && clients=0
+    cur=$(kubectl -n "$NS" get hpa orders -o jsonpath='{.status.currentMetrics[0].pods.current.averageValue}')
+    des=$(kubectl -n "$NS" get hpa orders -o jsonpath='{.status.desiredReplicas}')
+    ready=$(kubectl -n "$NS" get deploy orders -o jsonpath='{.status.readyReplicas}')
+    printf '%-6s %-10s %-9s %-8s %s\n' "${t}s" "$clients" "${cur:-?}" "${des:-?}" "${ready:-0}"
+    sleep 10
+  done
+  echo
+  echo "# HPA events:"
+  kubectl -n "$NS" get events --field-selector involvedObject.kind=HorizontalPodAutoscaler --sort-by=.lastTimestamp \
+    | awk 'NR>1 {$1=""; $2=""; $4=""; print}' | sed 's/^ *//'
+  echo
+  echo "# Load generator summary:"
+  kubectl -n "$NS" wait --for=jsonpath='{.status.phase}'=Succeeded pod/lg-hpa --timeout=60s > /dev/null
+  kubectl -n "$NS" logs lg-hpa | grep -E '^TOTAL|^LATENCY'
+} | tee "$OUT/hpa-custom-metric.txt"
+kubectl -n "$NS" logs lg-hpa > "$OUT/hpa-load.txt"
+kubectl -n "$NS" delete pod lg-hpa --wait=false > /dev/null
+kubectl -n "$NS" delete hpa orders > /dev/null
+kubectl -n "$NS" scale deploy/orders --replicas=2 > /dev/null