Find the fault.
Not just the drop.

What it does

Every string, compared against its own neighbours

A string is only meaningfully weak relative to the strings beside it, under the same sun, on the same box. That comparison is the whole detection method — and it is why soiling, shading and a failing module all surface long before they show up in plant output.

Self-hosting

Running on your own machine in ten minutes

Three containers — Postgres, the API, and nginx serving the dashboard. Published images, so nothing is compiled on the server. Any Linux box with Docker will do: Ubuntu 22.04 or newer, Debian 12, Amazon Linux 2023. Two vCPUs and 4 GB of RAM is comfortable.

1. 1

   Install Docker

   Skip if it is already there. Log out and back in afterwards so the group applies.

   Copy

   curl -fsSL https://get.docker.com | sudo sh
   sudo usermod -aG docker $USER

2. 2

   Write the secrets

   Three generated values. The cookie secret is what lets sessions survive a restart — the API refuses to start without it.

   Copy

   mkdir -p ~/solarops && cd ~/solarops
   cat > .env <<EOF
   POSTGRES_PASSWORD=$(openssl rand -hex 16)
   APP_DB_PASSWORD=$(openssl rand -hex 16)
   COOKIE_SECRET=$(openssl rand -hex 32)
   ADMIN_EMAIL=you@example.com
   ADMIN_PASSWORD=$(openssl rand -hex 12)
   EOF
   grep ADMIN .env        # note these down, it is your first login

3. 3

   Download the compose file

   One command. The API applies its own database migrations on start, so there is no separate setup step.

   Copy

   curl -fsSLO https://YOUR-SITE/docker-compose.yml

   What is in it

   Copy

   # SolarOps — self-hosted deployment.
   #
   #   curl -fsSLO https://YOUR-SITE/docker-compose.yml
   #   docker compose up -d
   #
   # Expects a .env beside it with POSTGRES_PASSWORD, APP_DB_PASSWORD,
   # COOKIE_SECRET, ADMIN_EMAIL and ADMIN_PASSWORD. See the Self-host section of
   # the site for a one-liner that generates them.
   
   services:
     db:
       image: postgres:16-alpine
       restart: unless-stopped
       environment:
         POSTGRES_USER: solarops
         POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:?set POSTGRES_PASSWORD in .env}
         POSTGRES_DB: solarops
       volumes:
         - pgdata:/var/lib/postgresql/data
       healthcheck:
         test: ["CMD-SHELL", "pg_isready -U solarops"]
         interval: 10s
         retries: 10
   
     backend:
       image: ghcr.io/solarenergyops/solarops/backend:latest
       restart: unless-stopped
       environment:
         NODE_ENV: production
         # The app connects as a restricted role; migrations connect as the owner.
         # Keeping them apart is what makes row-level security bite.
         DATABASE_URL: postgres://solarops_app:${APP_DB_PASSWORD:?set APP_DB_PASSWORD in .env}@db:5432/solarops
         MIGRATION_DATABASE_URL: postgres://solarops:${POSTGRES_PASSWORD}@db:5432/solarops
         APP_DB_USER: solarops_app
         APP_DB_PASSWORD: ${APP_DB_PASSWORD}
         COOKIE_SECRET: ${COOKIE_SECRET:?set COOKIE_SECRET in .env}
         # No TLS yet, so the session cookie cannot carry the Secure flag.
         # Put a certificate in front and delete this line.
         COOKIE_SECURE: "false"
         ADMIN_EMAIL: ${ADMIN_EMAIL:-}
         ADMIN_PASSWORD: ${ADMIN_PASSWORD:-}
       depends_on:
         db:
           condition: service_healthy
   
     frontend:
       image: ghcr.io/solarenergyops/solarops/frontend:latest
       restart: unless-stopped
       environment:
         API_UPSTREAM: http://backend:3000
       ports:
         - "80:8080"
       depends_on:
         - backend
   
   volumes:
     pgdata:

4. 4

   Start it

   First run pulls the images and migrates the schema. Under a minute after that.

   Copy

   docker compose up -d
   docker compose logs -f backend     # watch the migrations run