Performance Tuning

justapi serve --host 0.0.0.0 --port 8080 --workers 4

Each worker is a separate OS process. Workers share no state — use Redis/Postgres for shared data.

justapi serve --scale --min-workers 2 --max-workers 8 \
  --scale-high 1000 --scale-cooldown 30

app.set_database(
    "postgres://localhost/mydb",
    max_connections=20,               # Pool size
    request_acquire_timeout=1.0,      # Timeout for pool saturation
)

• Match pool size to expected concurrency
• Set request_acquire_timeout low (1-3s) for fast failure under load
• Use connection pooling behind the app for higher concurrency

app.run("0.0.0.0:8000", max_body_size=1024 * 1024)  # 1 MiB

Tight limits prevent memory exhaustion under load. Default is 50 MiB.

Set timeouts to bound worst-case latency:

justapi serve --timeout 30  # 30 second request timeout

Requests exceeding the timeout return 504 Gateway Timeout.

TLS adds approximately 10.7% overhead vs. plain HTTP. For maximum performance:

1. Terminate TLS at the edge proxy (nginx, Caddy, cloud LB)
2. Run JustAPI on a private loopback or Unix socket
3. Pass X-Forwarded-For and X-Forwarded-Proto from the proxy

Enable compression only when bandwidth is the bottleneck:

// In Cargo.toml features
justapi-core = { features = ["compression"] }

JustAPI automatically caches route lookups for high-traffic paths. No configuration needed — the cache is always active.

Maximum throughput comes from the native fast path:

@app.post("/fast", body_schema=Schema, native=True)
def fast_handler(request):
    return {"status": "ok"}  # 724k+ RPS

See Native Fast Path for details.

When running in Kubernetes, expect approximately 8.4% throughput degradation due to networking overhead. Mitigate by:

1. Using host networking mode
2. Setting resource requests/limits appropriately
3. Tuning kube-proxy to use iptables mode

# Warm up: 5 seconds
oha -z 5s http://localhost:8080/hello

# Benchmark: 30 seconds, 100 concurrent connections
oha -z 30s -c 100 http://localhost:8080/hello

# Results: p50, p95, p99 latency, RPS, peak RSS

Run minimum 3 runs per workload to establish distribution.

• Native Fast Path — 724k+ RPS execution
• Production Checklist — Production deployment settings
• Benchmarking Guide — Adding benchmarks