Skip to content

Persistence ​

njord uses Akka.Persistence to store scheduler state and forecast history. Two storage providers are supported.

Configuration ​

json
{
  "Njord": {
    "Persistence": {
      "Provider": "Sqlite"
    },
    "PersistencePath": "data/njord-journal.db"
  }
}

Providers ​

SQLite (default) ​

The default provider stores data in a local SQLite file. No external database required.

OptionDefaultDescription
Provider"Sqlite"Storage provider
PersistencePath"data/njord-journal.db"Path to the SQLite database file (relative to working directory)

In Docker, mount a volume at /app/data to persist the database across container restarts:

bash
docker run -v njord-data:/app/data ghcr.io/st0o0/njord:latest

PostgreSQL ​

For larger deployments or when you already run PostgreSQL, use the PostgreSql provider.

json
{
  "Njord": {
    "Persistence": {
      "Provider": "PostgreSql",
      "ConnectionString": "Host=db;Database=njord;Username=njord;Password=secret"
    }
  }
}
OptionDefaultDescription
Provider"PostgreSql"Storage provider
ConnectionStringnullPostgreSQL connection string (required when using PostgreSql)

WARNING

When using PostgreSQL, the ConnectionString is required. njord will refuse to start without it.

When is persistence needed? ​

Persistence is required for:

  • History enrichment: tracking forecast accuracy over time
  • Scheduler state: remembering poll cycle progress across restarts

If you disable the History enrichment and do not need restart resilience, persistence is optional but still recommended.

Open-Meteo data is licensed under CC BY 4.0.