durch
4e5ccf7926
Previously, when rotating API endpoints, only a single URL was provided to the HTTP client, defeating the purpose of having multiple URLs for resilience. Changes: - NymApiTopologyProvider now provides all URLs in rotated order when switching endpoints - NymApisClient similarly provides all URLs starting from the working endpoint - Added clarifying comments for broadcast/exhaustive query methods where single URLs are intentionally used - This enables the HTTP client's built-in failover mechanism while maintaining endpoint rotation behavior The fix ensures that if the primary endpoint fails, the client can automatically failover to alternative endpoints without manual intervention, improving overall network resilience.
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.