Merge pull request #604 from ClickHouse/revert-599-egork/incident-2025-08-31-services-in-aws-fail-after-provisioning

Revert "debug: diagnostics for incident 2025-08-31"

Author: alexey-milovidov

.github/workflows/clickhouse-cloud.yml

@@ -1,11 +1,6 @@
 name: "Run ClickBench on ClickHouse Cloud"
 on:
   workflow_dispatch:  # This allows manual trigger from the UI
-    inputs:
-      max_parallel:
-        description: "Max concurrent services (0 = no cap)"
-        required: false
-        default: "0"
   schedule:
     - cron: '10 10 * * *'
 
@@ -18,7 +13,6 @@ jobs:
     env:
       CI_COMMIT_MESSAGE: "[bot] update results for ClickHouse Cloud"
       CI_COMMIT_AUTHOR: github
-      MAX_PARALLEL: ${{ github.event.inputs.max_parallel || '0' }}
     steps:
     - uses: actions/checkout@v3
     - run: |
@@ -30,52 +24,21 @@ jobs:
           echo "Required secrets are not set. Skipping workflow."
           exit 0
         fi
+
         cd clickhouse-cloud
         curl https://clickhouse.com/ | sh
         sudo ./clickhouse install -y
-    - name: Install diagnostics tools
-      run: |
-        sudo apt-get update -y
-        sudo DEBIAN_FRONTEND=noninteractive apt-get install -y sysstat iproute2 traceroute moreutils procps dnsutils
-    - name: Start diagnostics collectors
-      working-directory: clickhouse-cloud
-      run: |
-        mkdir -p diag
-        # CPU usage per core every second
-        (mpstat -P ALL 1 > diag/mpstat.log 2>&1 &)
-        # System runq/mem/io snapshot every second
-        (vmstat 1 > diag/vmstat.log 2>&1 &)
-        # Network device and TCP stats every second (ignore if sar not available)
-        (sar -n DEV 1 > diag/sar_net_dev.log 2>&1 &) || true
-        (sar -n TCP,ETCP 1 > diag/sar_tcp.log 2>&1 &) || true
-        # Socket summary snapshot once per second
-        (for i in $(seq 1 3600); do date +%FT%T%z >> diag/ss.log; ss -s >> diag/ss.log; sleep 1; done) &
         
         bash combinations.sh
         bash collect-results.sh
         
-    - name: Upload diagnostics and raw run outputs
-      uses: actions/upload-artifact@v4
-      with:
-        name: clickhouse-cloud-diag-${{ github.run_id }}
-        path: |
-          clickhouse-cloud/diag/**
-          clickhouse-cloud/csp-*/netdebug.txt
-          clickhouse-cloud/csp-*/chclient_errors.log
-          clickhouse-cloud/csp-*/result
-          clickhouse-cloud/csp-*/state
-          clickhouse-cloud/results/*.json
-        if-no-files-found: ignore
-    - name: Commit results
-      working-directory: clickhouse-cloud
-      run: |
         git config --global user.name "${{ env.CI_COMMIT_AUTHOR }}"
         git config --global user.email "${{ env.CI_COMMIT_AUTHOR }}@users.noreply.github.com"
+        
         git pull
-        git add results/*.json || true
-        if git status --porcelain | grep -qE "^ M|^A |^\?\? "; then
+        git add results/*.json
+        if git status | grep -q modified
+        then
           git commit -m "${{ env.CI_COMMIT_MESSAGE }}"
           git push
-        else
-          echo "No changes to commit."
         fi

clickhouse-cloud/cloud-api.sh

@@ -86,27 +86,9 @@ done
 
 echo "Waiting for clickhouse-server to start"
 
-# Collect quick network/TLS diagnostics for this run
-{
-    echo "==== NET DEBUG $(date -Iseconds) ===="
-    echo "Host: $(hostname)"
-    echo "Kernel: $(uname -a)"
-    echo "ulimit -n: $(ulimit -n 2>/dev/null || echo n/a)"
-    echo "FQDN: ${FQDN}"
-    echo "IP route:"; ip route || true
-    echo "IP addr:"; ip -brief address || ip addr || true
-    echo "DNS resolution for ${FQDN}:"; getent hosts "${FQDN}" || nslookup "${FQDN}" 2>&1 || host "${FQDN}" 2>&1 || true
-    echo "Socket summary:"; ss -s || true
-    echo "Public IP (best-effort):"; curl -fsS --max-time 3 https://ifconfig.me || echo "n/a"
-    echo "OpenSSL s_client (short):"; timeout 10s openssl s_client -connect "${FQDN}:9440" -servername "${FQDN}" -brief </dev/null 2>&1 || true
-    echo "Traceroute TCP:9440:"; timeout 20s traceroute -T -p 9440 -n "${FQDN}" 2>&1 || traceroute -n "${FQDN}" 2>&1 || true
-} > "${TMPDIR}/netdebug.txt" 2>&1
-
 for i in {1..1000}
 do
-    if clickhouse-client --host "$FQDN" --password "$PASSWORD" --secure --query "SELECT 1" 1>/dev/null 2>>"${TMPDIR}/chclient_errors.log"; then
-        break
-    fi
+    clickhouse-client --host "$FQDN" --password "$PASSWORD" --secure --query "SELECT 1" && break
     sleep 1
     if [[ $i == 1000 ]]
     then

clickhouse-cloud/combinations.sh

@@ -8,22 +8,12 @@ PROVIDER=aws
 REGION='us-east-1'
 PARALLEL_REPLICA=false
 
-# Optional cap on parallel jobs; set MAX_PARALLEL=N in the environment to enable.
-throttle() {
-    local max=${MAX_PARALLEL:-0}
-    if [[ -n "$max" && "$max" =~ ^[0-9]+$ && $max -gt 0 ]]; then
-        while [[ $(jobs -rp | wc -l) -ge $max ]]; do
-            sleep 1
-        done
-    fi
-}
-
 for REPLICAS in 1
 do
     for MEMORY in 8 12
     do
         export PROVIDER REPLICAS REGION MEMORY PARALLEL_REPLICA
-    throttle; ./cloud-api.sh &
+        ./cloud-api.sh &
         sleep 10 # Prevent "Too many requests" to the API
     done
 done
@@ -33,7 +23,7 @@ do
     for MEMORY in 8 12 16 32 64 120 236
     do
         export PROVIDER REPLICAS REGION MEMORY PARALLEL_REPLICA
-    throttle; ./cloud-api.sh &
+        ./cloud-api.sh &
         sleep 10
     done
 done
@@ -46,7 +36,7 @@ do
     for MEMORY in 8 12
     do
         export PROVIDER REPLICAS REGION MEMORY PARALLEL_REPLICA
-    throttle; ./cloud-api.sh &
+        ./cloud-api.sh &
         sleep 10
     done
 done
@@ -56,7 +46,7 @@ do
     for MEMORY in 8 12 16 32 64 120 236
     do
         export PROVIDER REPLICAS REGION MEMORY PARALLEL_REPLICA
-    throttle; ./cloud-api.sh &
+        ./cloud-api.sh &
         sleep 10
     done
 done
@@ -69,7 +59,7 @@ do
     for MEMORY in 8 12
     do
         export PROVIDER REPLICAS REGION MEMORY PARALLEL_REPLICA
-    throttle; ./cloud-api.sh &
+        ./cloud-api.sh &
         sleep 10
     done
 done
@@ -79,7 +69,7 @@ do
     for MEMORY in 8 12 16 32 64 120
     do
         export PROVIDER REPLICAS REGION MEMORY PARALLEL_REPLICA
-    throttle; ./cloud-api.sh &
+        ./cloud-api.sh &
         sleep 10
     done
 done