How to Monitor Your Flowise Instance with Vigilmon

Flowise is an open-source drag-and-drop UI for building LLM-powered workflows with LangChain and LlamaIndex. Self-hosting Flowise gives you full control over your AI pipelines and API keys — but when Flowise goes down, your chatbots stop responding, your automation chains break, and users get no feedback. Vigilmon monitors every layer of your Flowise deployment and alerts you instantly when something fails.

What You'll Monitor

• The /api/v1/ping health endpoint (API liveness check)
• The Flowise web UI (drag-and-drop canvas interface)
• Background worker availability
• SSL certificate expiry

Prerequisites

• A self-hosted Flowise instance (Docker or npm/pm2)
• A free Vigilmon account
• Your Flowise URL (e.g. https://flowise.yourdomain.com)

Why Monitor Flowise?

Flowise runs as a Node.js server with an optional background queue for long-running LLM chains. A single deployment has multiple failure points:

• The Node.js process crashes and pm2 doesn't restart it
• The LLM API key quota is exhausted and prediction endpoints return errors
• The background worker queue backs up and new jobs silently time out
• Docker's memory limit kills the container mid-request

External monitoring from Vigilmon checks Flowise from outside your infrastructure — the way your users and downstream services experience it — and catches these failures before you do.

Step 1: Monitor the Ping Endpoint

Flowise ships with a built-in /api/v1/ping endpoint that confirms the server is alive and responsive.

1. Log in to vigilmon.online and click Add Monitor.
2. Set Type to HTTP / HTTPS.
3. Enter the URL: https://flowise.yourdomain.com/api/v1/ping
4. Set Check interval to 1 minute.
5. Set Expected status code to 200.
6. Under Expected body contains, enter pong — the endpoint returns { "ping": "pong" } when healthy.
7. Name it Flowise – API ping and click Save.

A healthy response:

{ "ping": "pong" }

If the server process dies or becomes unresponsive, this endpoint times out and Vigilmon triggers an alert.

Step 2: Monitor the Web UI

The Flowise canvas UI is what your team uses to build and debug LLM workflows. A broken UI means your team is blocked even when the API itself is responding.

1. Click Add Monitor.
2. Set Type to HTTP / HTTPS.
3. Enter: https://flowise.yourdomain.com
4. Set Expected status code to 200.
5. Under Expected body contains, enter Flowise — the page title confirms the React app is rendering.
6. Name it Flowise – Web UI and save.

Step 3: Monitor Background Workers

If you have Flowise configured with a queue-based worker (for parallel or long-running chains), the worker process can die independently from the main server. The main UI will appear healthy, but jobs submitted to the queue will silently stall.

The simplest way to monitor worker liveness is via a heartbeat:

1. In your Flowise worker startup script, add a periodic HTTP ping after each successful job cycle:

# cron every 2 minutes — add to crontab
*/2 * * * * /usr/bin/curl -s https://vigilmon.online/heartbeat/YOUR_HEARTBEAT_TOKEN

2. In Vigilmon, go to Heartbeat Monitors → Add Heartbeat.
3. Name it Flowise – Background Worker.
4. Set Expected interval to 2 minutes and Grace period to 1 minute.
5. Copy the generated URL and paste it into your worker script above.

If the worker stalls or dies, the ping stops arriving and Vigilmon alerts you after the grace period expires.

Step 4: SSL Certificate Alerts

Flowise is typically exposed behind a reverse proxy (nginx, Caddy, or Traefik). If your SSL certificate expires, all API calls from your chatbot integrations fail with TLS errors — and your users just see error messages.

1. Open any of the Flowise HTTP monitors you created.
2. Under SSL Certificate, enable Alert before expiry.
3. Set alerts at 30 days and 7 days before the certificate expires.
4. Apply the same setting to all monitors on the same domain.

Step 5: Set Up Alert Channels

1. In Vigilmon, go to Alert Channels → Add Channel.
2. Add your preferred channel — Email, Slack webhook, PagerDuty, or a generic webhook for your incident management tool.
3. For a Slack webhook, use:

{
  "text": "🚨 *{{monitor_name}}* is {{status}}!\nURL: {{url}}\nTime: {{timestamp}}"
}

4. Attach the channel to the ping monitor, the web UI monitor, and the heartbeat monitor.

Step 6: Set Response Time Thresholds

LLM inference is slow by nature, but your Flowise server itself should respond quickly to health checks and UI requests. A slow server is often a warning sign before a crash.

1. Open the ping monitor settings.
2. Under Response time threshold, set Warn to 500ms and Critical to 2000ms.
3. Open the web UI monitor and set Warn to 2000ms and Critical to 5000ms.

If Flowise's Node.js process is under heavy load or swapping memory, you'll see latency creep up before it fully fails.

Going Further

• Chatflow endpoint monitoring: If you have specific chatflows used in production, add individual monitors for POST /api/v1/prediction/{chatflow_id} using Vigilmon's HTTP monitor with a static request body. This confirms the whole inference pipeline is live, not just the server.
• API key rotation alerts: Set a calendar reminder aligned with your LLM API key expiry dates — combining this with Vigilmon's uptime data lets you correlate outages with key exhaustion.
• Status page: Share a Vigilmon public status page URL with your internal teams so they can check Flowise status without pinging on-call.

Your Flowise instance is now covered end-to-end: API liveness, web UI, background worker heartbeat, and SSL — with instant alerts before your LLM pipelines silently fail.