Merge pull request #1510 from lightpanda-io/wp/mrdimidium/ram-e2e-tests

Use cgroups for RAM mesurement

.github/workflows/e2e-test.yml

@@ -122,10 +122,19 @@ jobs:
     needs: zig-build-release
 
     env:
-      MAX_MEMORY: 26000
+      MAX_VmHWM: 26000 # 26MB (KB)
+      MAX_CG_PEAK: 6000 # 6MB (KB)
       MAX_AVG_DURATION: 17
       LIGHTPANDA_DISABLE_TELEMETRY: true
 
+      # How to give cgroups access to the user actions-runner on the host:
+      # $ sudo apt install cgroup-tools
+      # $ sudo chmod o+w /sys/fs/cgroup/cgroup.procs
+      # $ sudo mkdir -p /sys/fs/cgroup/actions-runner
+      # $ sudo chown -R actions-runner:actions-runner /sys/fs/cgroup/actions-runner
+      CG_ROOT: /sys/fs/cgroup
+      CG: actions-runner/lpd_${{ github.run_id }}_${{ github.run_attempt }}
+
     # use a self host runner.
     runs-on: lpd-bench-hetzner
     timeout-minutes: 15
@@ -150,10 +159,20 @@ jobs:
           go run ws/main.go & echo $! > WS.pid
           sleep 2
 
+      - name: run lightpanda in cgroup
+        run: |
+          if [ ! -f /sys/fs/cgroup/cgroup.controllers ]; then
+            echo "cgroup v2 not available: /sys/fs/cgroup/cgroup.controllers missing"
+            exit 1
+          fi
+
+          mkdir -p $CG_ROOT/$CG
+          cgexec -g memory:$CG ./lightpanda serve & echo $! > LPD.pid
+
+          sleep 2
+
       - name: run puppeteer
         run: |
-          ./lightpanda serve & echo $! > LPD.pid
-          sleep 2
           RUNS=100 npm run bench-puppeteer-cdp > puppeteer.out || exit 1
           cat /proc/`cat LPD.pid`/status |grep VmHWM|grep -oP '\d+' > LPD.VmHWM
           kill `cat LPD.pid`
@@ -161,11 +180,30 @@ jobs:
       - name: puppeteer result
         run: cat puppeteer.out
 
-      - name: memory regression
+      - name: cgroup memory regression
+        run: |
+          PID=$(cat LPD.pid)
+          while kill -0 $PID 2>/dev/null; do
+            sleep 1
+          done
+          if [ ! -f $CG_ROOT/$CG/memory.peak ]; then
+            echo "memory.peak not available in $CG"
+            exit 1
+          fi
+          PEAK_BYTES=$(cat $CG_ROOT/$CG/memory.peak)
+          PEAK_KB=$((PEAK_BYTES / 1024))
+          echo "memory.peak_bytes=$PEAK_BYTES"
+          echo "memory.peak_kb=$PEAK_KB"
+          test "$PEAK_KB" -le "$MAX_CG_PEAK"
+
+      - name: virtual memory regression
         run: |
           export LPD_VmHWM=`cat LPD.VmHWM`
           echo "Peak resident set size: $LPD_VmHWM"
-          test "$LPD_VmHWM" -le "$MAX_MEMORY"
+          test "$LPD_VmHWM" -le "$MAX_VmHWM"
+
+      - name: cleanup cgroup
+        run: rmdir $CG_ROOT/$CG
 
       - name: duration regression
         run: |
@@ -178,7 +216,8 @@ jobs:
           export AVG_DURATION=`cat puppeteer.out|grep 'avg run'|sed 's/avg run duration (ms) //'`
           export TOTAL_DURATION=`cat puppeteer.out|grep 'total duration'|sed 's/total duration (ms) //'`
           export LPD_VmHWM=`cat LPD.VmHWM`
-          echo "{\"duration_total\":${TOTAL_DURATION},\"duration_avg\":${AVG_DURATION},\"mem_peak\":${LPD_VmHWM}}" > bench.json
+          export LPD_CG_PEAK_KB=$(( $(cat $CG/memory.peak) / 1024 ))
+          echo "{\"duration_total\":${TOTAL_DURATION},\"duration_avg\":${AVG_DURATION},\"mem_peak\":${LPD_VmHWM},\"cg_mem_peak\":${LPD_MEM_CG_KB}}" > bench.json
           cat bench.json
 
       - name: run hyperfine