Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/FloxTBoTyy/BoardPulse-AI/llms.txt

Use this file to discover all available pages before exploring further.

Keeping an eye on your BoardPulse AI deployment helps you spot resource pressure early and confirm that all services are communicating correctly. This page covers the commands to monitor hardware usage, check service health, and diagnose the most common issues operators encounter.

​
GPU usage (NVIDIA)

When running Ollama with GPU support, verify that inference is hitting the GPU rather than falling back to CPU: 
docker compose exec ollama nvidia-smi
The output shows VRAM usage, GPU load percentage, and temperature for each card. Run this command during an active query to see utilization spike in real time.

​
CPU and RAM usage

To see a live view of processor and memory consumption across every running container: 
docker stats
This displays CPU percentage, memory usage, and network I/O for the api, admin, postgres, redis, openwebui, and ollama containers in a continuously refreshed table. Press Ctrl+C to exit.

​
API logs

Stream live output from the FastAPI backend to trace query handling, SQL generation, and errors: 
docker compose logs -f api
Replace api with any service name to tail its output: 
docker compose logs -f openwebui
docker compose logs -f ollama

​
Health check endpoint

The api service exposes a health check endpoint you can poll from any HTTP client: 
curl http://localhost:8000/api/v1/health
A healthy response returns: 
{
  "status": "ok",
  "environment": "development"
}
Use this endpoint in load balancers, uptime monitors, or deployment scripts to confirm the service is responsive before routing traffic.

​
Checking all service statuses

To see the status of every container in the stack at a glance: 
docker compose ps
All services should show a running status. A container in exited or restarting state indicates a problem — check its logs with docker compose logs -f <service> to find the error.

​
Admin panel connectivity check

The admin panel at http://localhost:3001 provides a visual check of your workspace configuration. Navigate to the Workspace default section to confirm that BoardPulse AI has successfully connected to your source database and detected the expected tables and columns. This is the fastest way to verify that DEFAULT_SOURCE_DATABASE_URL and DEFAULT_SOURCE_INCLUDE_TABLES are correct after changing your .env.

​
Common issues

The most common cause is an invalid or unreachable DATABASE_URL. Check that the postgres container is healthy before the api container attempts to connect:
docker compose ps postgres
docker compose logs api
Confirm that DATABASE_URL in your .env matches the credentials set in POSTGRES_USER, POSTGRES_PASSWORD, and POSTGRES_DB.
Verify two things: that OLLAMA_ENABLED=true is set in .env, and that the model has been pulled into the container.
docker compose exec ollama ollama list
If the model is missing from the list, pull it:
docker compose exec ollama ollama pull qwen3:8b
Also confirm that you started the stack with --profile local-models, since the ollama service does not start without it.
Export files are written to the path configured in EXPORTS_DIR (default: /tmp/boardpulse-exports). If exports silently fail, the directory may not be writable by the api process.Check the current value in your .env:
EXPORTS_DIR=/tmp/boardpulse-exports
If you have changed EXPORTS_DIR to a custom path, ensure the directory exists and the container user has write permission. Restart the api container after making changes:
docker compose up -d api

Build docs developers (and LLMs) love

Get started for freeTalk to us