Mark Sinclair
fb0b55d540
* ns-api: remove sqlite support ns-api: add env var to skip migrations for local dev ns-api: tidy up imports ns-api: fix deserialisation fo node descriptions update dockerfile update README fix up README and example env ns-api: bump major version to 4 ns-api: add more geoip data and new performance field in dvpn responses * ns-api: polyfill dVPN probe outcomes to make compatible with existing clients * Use explicit transaction for testrun status change (#6046) * Use explicit transaction for testrun status change * Improve run scripts * Skip locked rows * bump version 4.0.2 * Fix build.rs * Fix up .sqlx queries * Bump agent version and change dockerfile to run the agent in a loop * Make time between agents configurable by env var SLEEP_TIME * Update entrypoint.sh * Update Dockerfile with full path * Force bigint to avoid postgres numeric cast * Add override args to agent entry point, bump agent version and NS API version --------- Co-authored-by: Mark Sinclair <mmsinclair@users.noreply.github.com> Co-authored-by: dynco-nym <173912580+dynco-nym@users.noreply.github.com>
Nym Node Status API
The Node Status API serves information about individual nym-nodes in the Mixnet, such as which role they are operating in, statistics about them, services such as Network Requesters, as well as summaries of the state of the Mixnet.
We recommend that developers building applications such as explorers or analytics interfaces about the Mixnet run their own instance of the API, in order to promote a robust network of downstream services, and spread the load of API calls amongst as many endpoints as possible.
You can find build and operation instructions in the docs.
Database Support
The Node Status API supports both SQLite and PostgreSQL databases through Cargo feature flags:
- SQLite (default): Lightweight, file-based database suitable for development and small deployments
- PostgreSQL: Full-featured database recommended for production deployments
Building with Different Database Backends
# Build with SQLite (default)
cargo build --features sqlite --no-default-features
# Build with PostgreSQL
cargo build --features pg --no-default-features
Running Tests
# Test with SQLite
cargo test --features sqlite --no-default-features
# Test with PostgreSQL
make test-db # This sets up a test PostgreSQL instance
Development Commands
The project includes a Makefile with helpful commands for both database backends:
# Check code compilation
make check-sqlite # Check with SQLite
make check-pg # Check with PostgreSQL
# Run clippy linter
make clippy-sqlite # Lint with SQLite
make clippy-pg # Lint with PostgreSQL
make clippy # Run both
# PostgreSQL development
make dev-db # Start a PostgreSQL instance for development
make prepare-pg # Prepare SQLx offline cache for PostgreSQL
Implementation Details
The database abstraction is implemented using a query wrapper that automatically converts SQLite-style ? placeholders to PostgreSQL-style $1, $2, ... placeholders at runtime. This allows writing queries once using SQLite syntax while maintaining compatibility with both databases.
Key differences handled:
- Placeholder syntax (
?vs$1, $2, ...) - Type conversions (SQLite uses i64, PostgreSQL uses i32 for many fields)
- SQL dialect differences (e.g.,
INSERT OR IGNOREvsON CONFLICT DO NOTHING) - RETURNING clause behavior
For more details on PostgreSQL setup, see README_PG.md.