tutorial

Monitoring Continue.dev with Vigilmon: LLM Proxy Health, API Key Endpoints & AI Coding Uptime

How to monitor Continue.dev with Vigilmon — LLM proxy availability, model API endpoint health, SSL certificates, and alerting for AI coding assistant outages.

Continue.dev is an open-source AI coding assistant extension for VS Code and JetBrains that connects to any LLM — Claude, GPT-4, Llama, Ollama, and self-hosted models. Teams use Continue for inline code suggestions, multi-file context chat, and codebase Q&A. When the LLM backend or proxy that Continue routes requests through goes offline, developers lose autocomplete mid-session and chat queries return errors — breaking flow without a clear error message. Vigilmon gives you external visibility into the LLM API endpoints Continue depends on, the Continue Hub model proxy if you use it, and any self-hosted model backends your team runs.

What You'll Build

  • HTTP monitors for the LLM API endpoints Continue routes requests to
  • A monitor for the Continue Hub proxy if your team uses shared model access
  • SSL certificate monitoring for model API domains
  • An alerting setup that tells you which backend is down when Continue stops working

Prerequisites

  • Continue.dev installed in VS Code or JetBrains with a configured LLM provider
  • Knowledge of which model backends you use (Anthropic, OpenAI, Ollama, a custom proxy, or Continue Hub)
  • A free account at vigilmon.online

Step 1: Identify Your Continue.dev Backend Endpoints

Continue.dev's config.json (located at ~/.continue/config.json) defines which model APIs it calls. Open it to find your backends:

{
  "models": [
    {
      "title": "Claude 3.5 Sonnet",
      "provider": "anthropic",
      "model": "claude-sonnet-4-5",
      "apiBase": "https://api.anthropic.com"
    },
    {
      "title": "Local Llama",
      "provider": "ollama",
      "model": "llama3",
      "apiBase": "http://localhost:11434"
    }
  ]
}

Each apiBase is a candidate for monitoring. For cloud providers (Anthropic, OpenAI, Mistral), monitor the provider's API endpoint. For self-hosted backends (Ollama, LM Studio, vLLM), monitor your own server.

Step 2: Monitor the Anthropic API (if using Claude models)

If Continue is configured with Anthropic Claude models, monitor the Anthropic API health:

  1. Log in to Vigilmon → Add Monitor → HTTP.
  2. URL: https://api.anthropic.com.
  3. Check interval: 60 seconds.
  4. Response timeout: 15 seconds.
  5. Expected status: 200.
  6. Label: Anthropic API (Continue Claude backend).
  7. Click Save.

Step 3: Monitor the OpenAI API (if using GPT models)

For Continue configurations using GPT-4 or GPT-4o:

  1. Add Monitor → HTTP.
  2. URL: https://api.openai.com/v1/models.
  3. Check interval: 60 seconds.
  4. Response timeout: 15 seconds.
  5. Expected status: 200.
  6. Label: OpenAI API (Continue GPT backend).
  7. Click Save.

Step 4: Monitor a Self-Hosted Ollama Server

If your team runs Ollama locally or on a shared server for Continue, its health check endpoint is:

curl http://your-ollama-host:11434/api/tags

A healthy Ollama server returns a JSON list of installed models. Add a Vigilmon monitor:

  1. Add Monitor → HTTP.
  2. URL: http://your-ollama-host:11434/api/tags.
  3. Check interval: 60 seconds.
  4. Response timeout: 10 seconds.
  5. Expected status: 200.
  6. Keyword: models.
  7. Label: Ollama server (Continue local backend).
  8. Click Save.

If your Ollama server is only accessible on your LAN, Vigilmon's external checks won't reach it. In that case, use a local monitoring agent or rely on the cloud API monitors above for your primary alert path.

Step 5: Monitor the Continue Hub Proxy

Continue Hub provides shared model access for teams — a proxy that routes Continue requests to underlying LLMs with centralized billing and rate limit management. If your team uses Continue Hub, its API surface is at proxy.continue.dev:

curl -I https://proxy.continue.dev
  1. Add Monitor → HTTP.
  2. URL: https://proxy.continue.dev.
  3. Check interval: 2 minutes.
  4. Response timeout: 15 seconds.
  5. Expected status: 200.
  6. Label: Continue Hub proxy.
  7. Click Save.

Step 6: Monitor SSL Certificates

LLM API calls from Continue use TLS. An expired certificate on the model API domain causes Continue to fail with a TLS error that developers often misdiagnose as a network or firewall issue:

  1. Add Monitor → SSL Certificate.
  2. Domain: api.anthropic.com (or your primary model API domain).
  3. Alert when expiry is within: 30 days.
  4. Alert again: 14 days, 7 days, 3 days, 1 day.
  5. Click Save.

For self-hosted model backends with custom domains, add an SSL monitor for each:

  1. Add Monitor → SSL Certificate.
  2. Domain: your-ollama-host.example.com.
  3. Alert when expiry is within: 30 days.
  4. Click Save.

Step 7: Monitor a Custom LLM Proxy or Gateway

Many teams run a gateway (LiteLLM, OpenRouter, or a custom proxy) in front of multiple LLM providers to manage keys, rate limits, and model routing. Continue can point at these gateways via apiBase. Monitor your gateway's health endpoint:

# LiteLLM health endpoint
curl https://your-gateway.example.com/health
  1. Add Monitor → HTTP.
  2. URL: https://your-gateway.example.com/health.
  3. Check interval: 60 seconds.
  4. Expected status: 200.
  5. Keyword: healthy.
  6. Label: LLM gateway (Continue routing layer).
  7. Click Save.

Step 8: Configure Alerting

In Vigilmon under Settings → Notifications, configure your alert channels:

| Monitor | Trigger | Action | |---|---|---| | Anthropic API | Non-200 or timeout | Claude models in Continue unavailable; switch to local model | | OpenAI API | Non-200 or timeout | GPT models unavailable; check OpenAI status page | | Ollama server | Non-200 or keyword missing | Local model server down; restart Ollama or reconnect | | Continue Hub proxy | Non-200 | Shared team model access broken; alert dev team lead | | SSL certificate | < 30 days to expiry | TLS errors will follow; renew before expiry | | LLM gateway | Non-200 or keyword missing | All routes through gateway affected; check gateway logs |

Escalation policy: Route LLM backend alerts to the channel your engineering team watches. A Continue outage affects every developer using it simultaneously, so team-wide notification is appropriate.

Common Continue.dev Failure Modes and What Vigilmon Catches

| Scenario | Vigilmon monitor | |---|---| | Anthropic API outage | Anthropic monitor fires; Claude suggestions stop in Continue | | OpenAI rate limit exceeded | OpenAI monitor fires (429); developers see quota errors | | Ollama server crashes | Ollama monitor fires; local model suggestions fail | | Continue Hub proxy degraded | Hub monitor fires; team shared access broken | | SSL expiry on model API domain | SSL monitor alerts; Continue returns TLS errors | | LLM gateway misconfigured after deploy | Gateway health check fires; all models affected | | DNS failure for model API domain | Monitor fires from Vigilmon external check | | Firewall rule blocks model API egress | Monitor fires from Vigilmon (external) vs silent internally |

Continue.dev's effectiveness as a coding assistant depends entirely on the LLM backend being reachable. Vigilmon gives you external visibility into every backend Continue routes requests to — cloud model APIs, self-hosted Ollama servers, LLM gateways, and the Continue Hub proxy — so you know which layer failed when developers report that AI suggestions stopped working.

Start monitoring Continue.dev backends in under 5 minutes — register free at vigilmon.online.

Monitor your app with Vigilmon

Free plan — 5 monitors, no credit card required. Up and running in 60 seconds.

Start free →
← Back to all articles