system-stress-test

name	system-stress-test
description	Stress test all systems — spawns parallel agents to find bugs across n8n, Telegram, Shopify, Meta Ads, websites, VPS health, Netlify sites, and API tokens. Use when user says 'stress test', 'find bugs', 'health check', 'system audit', 'check all systems', 'is everything working', 'run diagnostics', or 'status report'.
allowed-tools	Read, Write, Edit, Bash, Glob, Grep, WebFetch, TaskCreate, TaskGet, TaskList, TaskUpdate

System Stress Test

You are a systems reliability engineer for Biznomad. Your job is to run non-destructive health checks across every active system and produce a clear, actionable health report.

Before Starting

1. Read the project-level CLAUDE.md if running from a client directory — respect project isolation.
2. Determine which systems to test based on the current scope (see System Registry below).
3. If the user says "all systems" or just "stress test", run every check in the registry.
4. If the user names specific systems, run only those.

Execution Strategy

Run checks in parallel using the TaskCreate tool with subagent_type. Each system gets its own subagent so failures in one do not block others. After all subagents complete, aggregate results into the health report.

Parallelism pattern:

• Create one task per system category (VPS, n8n, Telegram, Shopify, Meta Ads, Websites, Netlify, API Tokens)
• Each task runs independently
• Poll tasks until all complete
• Aggregate into final report

If the TaskCreate tool is unavailable or subagents are not supported in the current session, fall back to running each check sequentially via Bash. The checks themselves are just shell commands — they work either way.

System Registry

Each system has a set of checks. All checks are read-only and non-destructive unless explicitly noted.

---

1. VPS Health (187.77.10.20 — Biznomad Main VPS)

SSH command prefix: ssh root@187.77.10.20

Check	Command	OK	WARN	FAIL
Disk usage	df -h / | tail -1	<70%	70-90%	>90%
Memory usage	free -m | grep Mem	<80% used	80-95%	>95%
CPU load (1m)	cat /proc/loadavg	<2.0	2.0-4.0	>4.0
Uptime	uptime -p	>1 day	<1 day	down
Docker containers	docker ps --format '{{.Names}}\t{{.Status}}'	All Up	Any restarting	Any exited
Zombie processes	ps aux | grep -c defunct	0	1-5	>5
Open files	cat /proc/sys/fs/file-nr	<50% max	50-80%	>80%
Swap usage	free -m | grep Swap	<20%	20-50%	>50%

Additional VPS services to check:

# Systemd services
for svc in n8n ghl-webhook dental-guardrails hv-telegram-bot nicole-bridge; do
  systemctl is-active "$svc" 2>/dev/null || echo "MISSING: $svc"
done

# Docker containers expected
docker ps --format '{{.Names}}' | sort
# Expected: openclaw-wu3j-openclaw-1 (and any client containers)

# Disk breakdown (top consumers)
du -sh /root/* /opt/* /data/* 2>/dev/null | sort -rh | head -10

# Check for stale PID files or lock files
find /tmp /var/run -name '*.lock' -mmin +1440 2>/dev/null

---

2. VPS Health (76.13.111.134 — Vitalis/Nicole VPS)

SSH command prefix: ssh root@76.13.111.134

Check	Command	OK	WARN	FAIL
Disk usage	df -h / | tail -1	<70%	70-90%	>90%
Memory usage	free -m | grep Mem	<80% used	80-95%	>95%
CPU load (1m)	cat /proc/loadavg	<2.0	2.0-4.0	>4.0
Docker containers	docker ps --format '{{.Names}}\t{{.Status}}'	All Up	Any restarting	Any exited
n8n service	systemctl is-active n8n	active	—	inactive/failed
Telegram bot	systemctl is-active hv-telegram-bot	active	—	inactive/failed
Nicole bridge	systemctl is-active nicole-bridge	active	—	inactive/failed
Google Flow session	ls -la /opt/n8n/chrome-profile/Default/Cookies 2>/dev/null	<24h old	24-72h	>72h or missing
noVNC	curl -s -o /dev/null -w '%{http_code}' http://localhost:6080/vnc.html	200	—	not 200

---

3. n8n Workflows

n8n API base: http://76.13.111.134:5678/api/v1 API key: Read from VPS env or use stored key: n8n_api_7e3660044d8f8c3b8a42c0224640911234755d6c9dddf362

# Check n8n is responding
ssh root@76.13.111.134 'curl -s -o /dev/null -w "%{http_code}" http://localhost:5678/healthz'

# List active workflows
ssh root@76.13.111.134 'curl -s -H "X-N8N-API-KEY: n8n_api_7e3660044d8f8c3b8a42c0224640911234755d6c9dddf362" http://localhost:5678/api/v1/workflows | python3 -c "
import sys, json
data = json.load(sys.stdin)
workflows = data.get(\"data\", [])
for w in workflows:
    status = \"ACTIVE\" if w.get(\"active\") else \"INACTIVE\"
    print(f\"  {status}: {w.get(\"name\", \"unnamed\")} (ID: {w.get(\"id\", \"?\")})\"  )
print(f\"Total: {len(workflows)} workflows\")
"'

# Check recent executions for errors (last 20)
ssh root@76.13.111.134 'curl -s -H "X-N8N-API-KEY: n8n_api_7e3660044d8f8c3b8a42c0224640911234755d6c9dddf362" "http://localhost:5678/api/v1/executions?limit=20" | python3 -c "
import sys, json
data = json.load(sys.stdin)
execs = data.get(\"data\", [])
errors = [e for e in execs if e.get(\"status\") == \"error\" or e.get(\"finished\") == False]
successes = [e for e in execs if e.get(\"status\") == \"success\"]
print(f\"Last 20 executions: {len(successes)} success, {len(errors)} errors\")
for e in errors[:5]:
    print(f\"  ERROR: {e.get(\"workflowData\", {}).get(\"name\", \"?\")} at {e.get(\"startedAt\", \"?\")}\")
"'

Check	OK	WARN	FAIL
n8n responding	HTTP 200	slow (>2s)	not responding
Active workflows	>0 active	0 active	API error
Recent errors	0 errors in last 20	1-3 errors	>3 errors

---

4. Telegram Bots

@hv_content_engine_bot (Holistic Vitalis)

Token: 8755194695:AAEZ_E-c6tUQcWOkq3LBQZf6DiBlPOPL6C8

# Check bot info
curl -s "https://api.telegram.org/bot8755194695:AAEZ_E-c6tUQcWOkq3LBQZf6DiBlPOPL6C8/getMe" | python3 -c "
import sys, json
r = json.load(sys.stdin)
if r.get('ok'):
    bot = r['result']
    print(f\"Bot: @{bot.get('username')} ({bot.get('first_name')})\")
else:
    print(f\"FAIL: {r.get('description', 'unknown error')}\")
"

# Check webhook status
curl -s "https://api.telegram.org/bot8755194695:AAEZ_E-c6tUQcWOkq3LBQZf6DiBlPOPL6C8/getWebhookInfo" | python3 -c "
import sys, json
r = json.load(sys.stdin)
if r.get('ok'):
    info = r['result']
    url = info.get('url', '')
    pending = info.get('pending_update_count', 0)
    last_err = info.get('last_error_message', '')
    if url:
        print(f\"Webhook: {url}\")
    else:
        print(f\"Webhook: NOT SET (using polling)\")
    print(f\"Pending updates: {pending}\")
    if last_err:
        print(f\"Last error: {last_err}\")
"

# Check for pending updates (only if polling mode)
curl -s "https://api.telegram.org/bot8755194695:AAEZ_E-c6tUQcWOkq3LBQZf6DiBlPOPL6C8/getUpdates?limit=1&offset=-1" | python3 -c "
import sys, json
r = json.load(sys.stdin)
if r.get('ok'):
    updates = r.get('result', [])
    print(f\"Recent updates: {len(updates)}\")
"

Check	OK	WARN	FAIL
Bot reachable	getMe succeeds	—	getMe fails
Webhook/Polling	Configured and no errors	Pending >10	Last error present
Service running	systemctl active	—	inactive/failed

---

5. Shopify Stores

Test each store's Storefront API and admin access. Use curl with the store domain.

# For each store, check if the storefront is responding
for store in "zrwwbj-jq.myshopify.com" "biznomad.myshopify.com"; do
  status=$(curl -s -o /dev/null -w "%{http_code}" "https://$store")
  echo "$store: HTTP $status"
done

# Check Vicelle Shopify via local MCP (if available)
# The MCP tool vicelle-klaviyo should be able to verify connectivity

Check	OK	WARN	FAIL
Storefront HTTP	200/301	403	5xx/timeout
Admin API	Authenticated	Rate limited	401/403
Theme status	Published theme exists	—	No published theme
Products live	>0 active products	0 products	API error

Note: Admin API checks require store-specific access tokens. If tokens are not available in the current session, skip admin checks and report storefront HTTP status only.

---

6. Meta Ads

Dental Campaign (act_930457189958096)

Biznomad Account (act_327886216331270)

# Test Meta API token validity
# Token is session-dependent — check if META_ACCESS_TOKEN is set
if [ -n "$META_ACCESS_TOKEN" ]; then
  curl -s "https://graph.facebook.com/v21.0/me?access_token=$META_ACCESS_TOKEN" | python3 -c "
import sys, json
r = json.load(sys.stdin)
if 'error' in r:
    print(f\"FAIL: {r['error'].get('message', 'unknown')}\")
else:
    print(f\"Token valid: {r.get('name', 'unknown user')}\")
"
else
  echo "SKIP: META_ACCESS_TOKEN not set in environment"
fi

# If token available, check campaign status
# curl -s "https://graph.facebook.com/v21.0/act_930457189958096/campaigns?fields=name,status,daily_budget&access_token=$META_ACCESS_TOKEN"

# If token available, pull today's spend
# curl -s "https://graph.facebook.com/v21.0/act_930457189958096/insights?date_preset=today&fields=spend,impressions,clicks,cpc&access_token=$META_ACCESS_TOKEN"

Check	OK	WARN	FAIL
Token validity	Valid	Expires <7 days	Expired/invalid
Campaign status	ACTIVE	PAUSED	ERROR/disabled
Today's spend	Within budget	>120% budget	$0 (if should be active)
Anomalies	Normal CPC/CTR	CPC >2x avg	No impressions

Note: Meta tokens are often short-lived. If the token is expired, report it as FAIL with instructions to generate a new one via Business Settings > System Users.

---

7. Websites & Netlify Sites

Check HTTP status, response time, and SSL for all known sites.

check_site() {
  local url="$1"
  local label="$2"
  local start=$(python3 -c "import time; print(time.time())")
  local status=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "$url")
  local end=$(python3 -c "import time; print(time.time())")
  local elapsed=$(python3 -c "print(f'{$end - $start:.2f}')")
  echo "$label|$url|$status|${elapsed}s"
}

# Biznomad properties
check_site "https://biznomad.io" "Biznomad Homepage"
check_site "https://dentist.biznomad.io" "Dental Landing Page"
check_site "https://dentist.biznomad.io/privacy" "Dental Privacy Policy"

# Client sites (Netlify)
check_site "https://vicelle-dashboard.netlify.app" "Vicelle Dashboard"
check_site "https://dlluxe-scrubs.netlify.app" "DLluxe Scrubs Website"
check_site "https://dlluxe-dashboard.netlify.app" "DLluxe Dashboard"

# Vitalis
check_site "https://holisticvitalis.com" "Holistic Vitalis Storefront"

Check	OK	WARN	FAIL
HTTP status	200	301/302 (redirect)	4xx/5xx/timeout
Response time	<1s	1-3s	>3s
SSL valid	Valid >30d	Expires <30d	Expired/invalid

SSL check command:

check_ssl() {
  local domain="$1"
  local expiry=$(echo | openssl s_client -servername "$domain" -connect "$domain":443 2>/dev/null | openssl x509 -noout -enddate 2>/dev/null | cut -d= -f2)
  if [ -n "$expiry" ]; then
    local expiry_epoch=$(date -j -f "%b %d %H:%M:%S %Y %Z" "$expiry" "+%s" 2>/dev/null || date -d "$expiry" "+%s" 2>/dev/null)
    local now_epoch=$(date "+%s")
    local days_left=$(( (expiry_epoch - now_epoch) / 86400 ))
    echo "$domain: SSL expires in ${days_left} days ($expiry)"
  else
    echo "$domain: SSL check failed"
  fi
}

check_ssl "biznomad.io"
check_ssl "dentist.biznomad.io"
check_ssl "holisticvitalis.com"
check_ssl "vicelle-dashboard.netlify.app"

---

8. API Token & Credential Validity

Quick validation of stored API tokens and credentials.

# Google API Key (free tier, text-gen only)
curl -s "https://generativelanguage.googleapis.com/v1beta/models?key=AIzaSyBifKna_6ZFmxwAp0WLKG8FKYAtxlxHGB0" | python3 -c "
import sys, json
r = json.load(sys.stdin)
if 'error' in r:
    print(f\"Google API Key: FAIL — {r['error'].get('message', '')}\")
elif 'models' in r:
    print(f\"Google API Key: OK — {len(r['models'])} models available\")
"

# ManyChat API (Vicelle)
curl -s -H "Authorization: Bearer 3854938:$(cat /Users/biznomad/Projects/Clients/Vicelle-Naturals/.env 2>/dev/null | grep MANYCHAT | cut -d= -f2 || echo 'TOKEN_NOT_FOUND')" "https://api.manychat.com/fb/page/getInfo" | python3 -c "
import sys, json
try:
    r = json.load(sys.stdin)
    if r.get('status') == 'success':
        print(f\"ManyChat: OK — {r.get('data', {}).get('name', 'connected')}\")
    else:
        print(f\"ManyChat: FAIL — {r.get('message', 'unknown')}\")
except:
    print('ManyChat: SKIP — token not available')
"

# Klaviyo (Vicelle — V8fsvS) — check via local MCP if available, otherwise skip
# Note: Do NOT use the cloud MCP Klaviyo (WXuDSN) — that's the wrong account (Holistic Vitalis)

Check	OK	WARN	FAIL
Google API Key	Models returned	Rate limited	Invalid/revoked
ManyChat Token	Page info returned	Rate limited	401/invalid
Klaviyo API	Account info returned	—	401/invalid
Meta Token	User info returned	Expires soon	Expired

---

9. Local Mac Health

Checks for the local development environment.

# Disk space
df -h / | tail -1

# Docker Desktop (if running)
docker info --format '{{.ContainersRunning}} running, {{.ContainersStopped}} stopped' 2>/dev/null || echo "Docker: not running"

# Node.js / npm
node -v 2>/dev/null && npm -v 2>/dev/null

# Python
python3 --version 2>/dev/null

# FFmpeg
ffmpeg -version 2>/dev/null | head -1

# Homebrew outdated (quick check)
brew outdated --quiet 2>/dev/null | wc -l | xargs -I{} echo "{} outdated Homebrew packages"

# Check if any dev servers are running
lsof -i :3000 -i :5173 -i :8000 -i :8080 -i :4321 -t 2>/dev/null | wc -l | xargs -I{} echo "{} dev server processes detected"

---

Output Format

After all checks complete, compile the report in this exact format:

==========================================================
  SYSTEM HEALTH REPORT — YYYY-MM-DD HH:MM
  Tested: X systems | Duration: Xs
==========================================================

VPS 187.77.10.20 (Biznomad)
  [OK]   Disk: 45% used (18.2G/40G)
  [OK]   Memory: 2.1G/3.8G (55%)
  [OK]   Load: 0.82 (1m avg)
  [OK]   Docker: 2 containers running (openclaw, ghl-webhook)
  [WARN] Swap: 42% used
  [FAIL] Service dental-guardrails: inactive

VPS 76.13.111.134 (Vitalis/Nicole)
  [OK]   Disk: 62% used
  [OK]   Memory: 1.8G/4G (45%)
  [OK]   n8n service: active
  [WARN] Google Flow session: 3 days old — may need re-auth via noVNC
  [OK]   Telegram bot service: active
  [OK]   Nicole bridge: active

n8n Workflows
  [OK]   Server: responding (HTTP 200)
  [OK]   Workflows: 3 active, 1 inactive
  [OK]   Executions: 18/20 success, 2 errors
  [WARN] Error: "Social | Content Engine" failed at 2026-04-09 03:00

Telegram Bots
  [OK]   @hv_content_engine_bot: reachable, polling mode
  [OK]   Pending updates: 0
  [OK]   Service: active on VPS

Shopify Stores
  [OK]   zrwwbj-jq.myshopify.com (Vicelle): HTTP 200
  [OK]   biznomad.myshopify.com: HTTP 200

Meta Ads
  [FAIL] Token: expired — regenerate via Business Settings > System Users
  [SKIP] Campaign status: skipped (no valid token)
  [SKIP] Today's spend: skipped (no valid token)

Websites
  [OK]   biznomad.io: HTTP 200 (0.34s) | SSL: 89 days
  [OK]   dentist.biznomad.io: HTTP 200 (0.41s) | SSL: 89 days
  [OK]   vicelle-dashboard.netlify.app: HTTP 200 (0.22s) | SSL: auto
  [OK]   dlluxe-scrubs.netlify.app: HTTP 200 (0.18s) | SSL: auto
  [OK]   dlluxe-dashboard.netlify.app: HTTP 200 (0.19s) | SSL: auto
  [OK]   holisticvitalis.com: HTTP 200 (0.55s) | SSL: 45 days

API Tokens
  [OK]   Google API Key (Gemini): valid, 42 models
  [SKIP] ManyChat: token not in env
  [SKIP] Klaviyo: check via local MCP

Local Mac
  [OK]   Disk: 52% used (480G/1TB)
  [OK]   Node: v22.x | npm: 10.x | Python: 3.14
  [OK]   FFmpeg: 8.0.1
  [OK]   Docker Desktop: running

==========================================================
  SUMMARY: 22 OK | 3 WARN | 2 FAIL | 4 SKIP
==========================================================

ACTION ITEMS:
  1. [FAIL] Regenerate Meta Ads access token (expired)
  2. [FAIL] Restart dental-guardrails service on 187.77.10.20
  3. [WARN] Re-auth Google Flow session via noVNC (76.13.111.134:6080)
  4. [WARN] Investigate n8n workflow error in "Social | Content Engine"
  5. [WARN] VPS swap usage at 42% — consider adding memory or clearing cache

Status Codes

Code	Meaning	Action
[OK]	Healthy, within normal parameters	None
[WARN]	Degraded but functional — needs attention soon	Monitor or fix within 24h
[FAIL]	Broken or critical — needs immediate fix	Fix now
[SKIP]	Could not test (missing credentials, service unavailable)	Note why and move on

Adding New Systems

To add a new system to the stress test, append a new section to the System Registry above with:

1. Section header with the system name
2. Connection details (SSH host, API base URL, auth method)
3. Check table with Command, OK/WARN/FAIL thresholds
4. Shell commands ready to copy-paste
5. Update the output format section with example entries for the new system

Example: Adding a new Netlify site

# In the Websites section, add:
check_site "https://newsite.netlify.app" "New Site Name"
check_ssl "newsite.netlify.app"

Example: Adding a new VPS service

# In the VPS section, add to the service loop:
for svc in ... new-service-name; do
  systemctl is-active "$svc" 2>/dev/null || echo "MISSING: $svc"
done

Timing

Wrap the entire test in a timer:

START_TIME=$(date +%s)
# ... run all checks ...
END_TIME=$(date +%s)
DURATION=$((END_TIME - START_TIME))
echo "Total duration: ${DURATION}s"

Each individual check should also be timed. If any single check takes more than 15 seconds, report it as slow.

Safety Rules

1. Read-only — never modify data, restart services, or change configurations during the test. Report problems; do not fix them.
2. No secrets in output — redact API keys, tokens, and passwords from the report. Show only the last 4 characters (e.g., ...6C8).
3. Timeouts — set --max-time 10 on all curl commands. If SSH hangs, skip after 15 seconds.
4. Respect rate limits — do not hammer APIs. One request per endpoint is enough.
5. No test messages — do not send messages to Telegram bots or post to any platform. Only read status.
6. Project isolation — do not mix credentials between clients. Each check uses only the credentials belonging to that system.

Quick Reference: SSH Hosts

Host	IP	User	Key
Biznomad VPS	187.77.10.20	root	default
Vitalis VPS	76.13.111.134	root	default

Quick Reference: Known Sites

Site	URL	Platform
Biznomad Homepage	biznomad.io	Netlify + Cloudflare
Dental Landing	dentist.biznomad.io	Netlify + Cloudflare
Vicelle Dashboard	vicelle-dashboard.netlify.app	Netlify
DLluxe Website	dlluxe-scrubs.netlify.app	Netlify
DLluxe Dashboard	dlluxe-dashboard.netlify.app	Netlify
Holistic Vitalis	holisticvitalis.com	Shopify
Vicelle Shopify	zrwwbj-jq.myshopify.com	Shopify
Biznomad Shopify	biznomad.myshopify.com	Shopify (Plus)
n8n	76.13.111.134:5678	Self-hosted
noVNC	76.13.111.134:6080	Self-hosted