frameworks/node: add benchmark implementation

[nigrosimone]

Node was on the board only through frameworks (express, fastify, koa, nestjs, node-h3, hono-node), with no bare node:http line to read them against. bun, deno and go-stdlib already have one.

One file, no dependencies: http.createServer, routing by string comparison on req.url, query parsed by hand, and zlib.gzipSync on /json because node:http negotiates nothing.

node:cluster forks one worker per core, but each worker binds 8080 with reusePort, so the kernel spreads the accepts the way bun and deno do, not the cluster round robin.

Profiles are the same six as bun, deno and go-stdlib. Checked with the image on 4 cpus: 4 workers, 4 LISTEN sockets on 8080, and every endpoint the profiles use.

[MDA2AV]

/benchmark --save

[github-actions]

👋 Benchmark request received. A collaborator will review and approve the run.

[github-actions]

Benchmark Results

Framework: node | Test: all tests

Test	Conn	RPS	CPU	Mem	Δ RPS	Δ Mem
baseline	512	907,802	6679.8%	1.5GiB	~0%	~0%
baseline	4096	892,087	6553.9%	1.6GiB	~0%	~0%
pipelined	512	1,848,457	6669.2%	2.3GiB	~0%	~0%
pipelined	4096	1,837,282	6553.1%	2.5GiB	~0%	~0%
limited-conn	512	747,342	6187.4%	5.3GiB	~0%	~0%
limited-conn	4096	768,440	6415.9%	5.4GiB	~0%	~0%
json	4096	713,493	6393.6%	5.0GiB	~0%	~0%
json-comp	512	281,475	5817.4%	8.9GiB	~0%	~0%
json-comp	4096	289,309	6307.9%	9.0GiB	~0%	~0%
json-comp	16384	285,229	6180.5%	9.5GiB	~0%	~0%
upload	32	2,058	2462.3%	5.3GiB	~0%	~0%
upload	256	2,156	5931.9%	5.6GiB	~0%	~0%

Full log

[info] server ready

[run 1/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     32 (0/thread)
  Pipeline:  1
  Req/conn:  5
  Templates: 4
  Expected:  200
  Duration:  5s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   15.48ms   8.97ms   37.30ms   78.90ms   123.00ms

  10292 requests in 5.00s, 10292 responses
  Throughput: 2.06K req/s
  Bandwidth:  340.07KB/s
  Status codes: 2xx=10292, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 10292 / 10292 responses (100.0%)
  Reconnects: 2056
  Per-template: 2558,2551,2596,2587
  Per-template-ok: 2558,2551,2596,2587
[info] CPU 2462.3% | Mem 5.3GiB

[run 2/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     32 (0/thread)
  Pipeline:  1
  Req/conn:  5
  Templates: 4
  Expected:  200
  Duration:  5s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   15.56ms   8.88ms   37.20ms   83.50ms   130.60ms

  10266 requests in 5.00s, 10266 responses
  Throughput: 2.05K req/s
  Bandwidth:  339.22KB/s
  Status codes: 2xx=10266, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 10266 / 10266 responses (100.0%)
  Reconnects: 2055
  Per-template: 2562,2554,2564,2586
  Per-template-ok: 2562,2554,2564,2586
[info] CPU 2654.0% | Mem 5.4GiB

[run 3/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     32 (0/thread)
  Pipeline:  1
  Req/conn:  5
  Templates: 4
  Expected:  200
  Duration:  5s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   15.60ms   8.77ms   37.90ms   91.30ms   125.60ms

  10234 requests in 5.00s, 10234 responses
  Throughput: 2.04K req/s
  Bandwidth:  338.17KB/s
  Status codes: 2xx=10234, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 10234 / 10234 responses (100.0%)
  Reconnects: 2047
  Per-template: 2553,2544,2558,2579
  Per-template-ok: 2553,2544,2558,2579
[info] CPU 2504.0% | Mem 5.4GiB

=== Best: 2058 req/s (CPU: 2462.3%, Mem: 5.3GiB) ===
[info] input BW: 16.32GB/s (avg template: 8516680 bytes)
[info] saved results/upload/32/node.json
httparena-bench-node
httparena-bench-node

==============================================
=== node / upload / 256c (tool=gcannon) ===
==============================================
[info] waiting for server...
[info] server ready

[run 1/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     256 (4/thread)
  Pipeline:  1
  Req/conn:  5
  Templates: 4
  Expected:  200
  Duration:  5s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   115.53ms   63.00ms   306.60ms   656.70ms   940.00ms

  10785 requests in 5.00s, 10783 responses
  Throughput: 2.15K req/s
  Bandwidth:  356.28KB/s
  Status codes: 2xx=10783, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 10783 / 10783 responses (100.0%)
  Reconnects: 2115
  Per-template: 2459,2808,2899,2617
  Per-template-ok: 2459,2808,2899,2617
[info] CPU 5931.9% | Mem 5.6GiB

[run 2/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     256 (4/thread)
  Pipeline:  1
  Req/conn:  5
  Templates: 4
  Expected:  200
  Duration:  5s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   116.72ms   63.10ms   305.30ms   709.00ms   975.30ms

  10724 requests in 5.00s, 10723 responses
  Throughput: 2.14K req/s
  Bandwidth:  354.31KB/s
  Status codes: 2xx=10723, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 10722 / 10723 responses (100.0%)
  Reconnects: 2106
  Per-template: 2414,2786,2905,2617
  Per-template-ok: 2414,2786,2905,2617
[info] CPU 6134.0% | Mem 5.7GiB

[run 3/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     256 (4/thread)
  Pipeline:  1
  Req/conn:  5
  Templates: 4
  Expected:  200
  Duration:  5s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   116.09ms   61.10ms   300.80ms   730.00ms    1.06s

  10771 requests in 5.00s, 10771 responses
  Throughput: 2.15K req/s
  Bandwidth:  355.91KB/s
  Status codes: 2xx=10771, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 10767 / 10771 responses (100.0%)
  Reconnects: 2110
  Per-template: 2449,2791,2898,2629
  Per-template-ok: 2449,2791,2898,2629
[info] CPU 5941.2% | Mem 5.7GiB

=== Best: 2156 req/s (CPU: 5931.9%, Mem: 5.6GiB) ===
[info] input BW: 17.10GB/s (avg template: 8516680 bytes)
[info] saved results/upload/256/node.json
httparena-bench-node
httparena-bench-node
[info] skip: node does not subscribe to api-4
[info] skip: node does not subscribe to api-16
[info] skip: node does not subscribe to static
[info] skip: node does not subscribe to static-tls
[info] skip: node does not subscribe to async-db
[info] skip: node does not subscribe to crud
[info] skip: node does not subscribe to fortunes
[info] skip: node does not subscribe to baseline-h2
[info] skip: node does not subscribe to static-h2
[info] skip: node does not subscribe to baseline-h2c
[info] skip: node does not subscribe to json-h2c
[info] skip: node does not subscribe to baseline-h3
[info] skip: node does not subscribe to static-h3
[info] skip: node does not subscribe to gateway-64
[info] skip: node does not subscribe to gateway-h3
[info] skip: node does not subscribe to production-stack
[info] skip: node does not subscribe to unary-grpc
[info] skip: node does not subscribe to unary-grpc-tls
[info] skip: node does not subscribe to stream-grpc
[info] skip: node does not subscribe to stream-grpc-tls
[info] skip: node does not subscribe to echo-ws
[info] skip: node does not subscribe to echo-ws-pipeline
[info] skip: node does not subscribe to echo-ws-limited
[info] rebuilding site/data/*.json
[purged stale] zix-grpc.json (zix-grpc)
[purged stale] zix-ws.json (zix-ws)
[updated] /home/diogo/actions-runner/_work/HttpArena/HttpArena/site/data/frameworks.json
[updated] /home/diogo/actions-runner/_work/HttpArena/HttpArena/site/data/results/node.json - 12 new, 12 total
[updated] /home/diogo/actions-runner/_work/HttpArena/HttpArena/site/data/current.json
[info] done
[info] restoring loopback MTU to 65536