fastapi: emit a single Server header and stop gzip-inflating tiny bodies

[TOKUJI]

Fixes #1052.

• Server header: drop the BaseHTTPMiddleware-based ServerHeaderMiddleware and pass --no-server-header --header Server:FastAPI to uvicorn instead. Responses previously carried both server: uvicorn and server: FastAPI; they now carry exactly one server: FastAPI, at zero per-request cost.
• GZip: minimum_size=1 → 1000 — the value in FastAPI's own GZipMiddleware docs example; compresslevel=5 unchanged. Happy to switch to Starlette's default (500) if you prefer.

Verified by building the entry's Dockerfile (uvicorn 0.40.0) and probing the container:

• /baseline11?a=13&b=42 → 200, single server: FastAPI, and no Content-Encoding even with Accept-Encoding: gzip (the 2-byte body is no longer inflated)
• /json/50?m=1 with Accept-Encoding: gzip, br → Content-Encoding: gzip (7602 B → 423 B); no Content-Encoding without Accept-Encoding — the json-comp validation checks still pass

🤖 Generated with Claude Code

[Kaliumhexacyanoferrat]

/benchmark -f fastapi --save

[github-actions]

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

[github-actions]

Benchmark Results

Framework: fastapi | Test: all tests

Test	Conn	RPS	CPU	Mem	Δ RPS	Δ Mem
baseline	512	320,447	6505.3%	4.5GiB	+129.2%	-4.3%
baseline	4096	314,605	6506.2%	4.6GiB	+154.7%	-13.2%
pipelined	512	364,432	6524.6%	4.5GiB	+150.0%	-6.2%
pipelined	4096	359,847	6524.2%	4.7GiB	+193.5%	-14.5%
limited-conn	512	284,181	6508.9%	5.7GiB	+119.5%	+11.8%
limited-conn	4096	286,820	6496.5%	5.8GiB	+146.6%	+5.5%
json	4096	157,204	6468.8%	4.8GiB	+83.6%	-9.4%
json-comp	512	93,695	6450.8%	4.7GiB	+45.2%	-2.1%
json-comp	4096	93,953	6445.4%	4.7GiB	+50.6%	-13.0%
json-comp	16384	92,206	6450.5%	4.9GiB	+65.4%	-26.9%
json-tls	4096	135,982	6176.2%	4.8GiB	+78.9%	-15.8%
upload	32	2,125	2532.7%	4.6GiB	+37.9%	~0%
upload	256	1,536	6271.6%	4.8GiB	+6.7%	-4.0%
api-4	256	8,449	401.4%	494MiB	+10.7%	-0.2%
api-16	1024	30,215	1616.1%	1.7GiB	+61.8%	+13.3%
static	1024	22,055	6702.8%	5.3GiB	+12.9%	-7.0%
static	4096	20,911	6765.1%	6.8GiB	+16.9%	-4.2%
static	6800	20,315	6712.9%	7.9GiB	+17.0%	-3.7%
async-db	1024	76,351	6248.4%	4.8GiB	+39.9%	-4.0%

Full log

[run 1/3]
Running 5s test @ http://localhost:8080
  64 threads and 4096 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   191.84ms   96.00ms 978.58ms   80.71%
    Req/Sec   326.92    109.72   760.00     81.88%
  98065 requests in 5.12s, 1.72GB read
Requests/sec:  19137.45
Transfer/sec:    343.34MB
[info] CPU 6618.8% | Mem 6.6GiB

[run 2/3]
Running 5s test @ http://localhost:8080
  64 threads and 4096 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   187.91ms   64.01ms 472.21ms   81.41%
    Req/Sec   335.26     47.70   580.00     78.79%
  106844 requests in 5.11s, 1.87GB read
Requests/sec:  20911.57
Transfer/sec:    375.36MB
[info] CPU 6765.1% | Mem 6.8GiB

[run 3/3]
Running 5s test @ http://localhost:8080
  64 threads and 4096 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   189.01ms   65.52ms 495.68ms   80.92%
    Req/Sec   333.20     46.57   575.00     75.96%
  106186 requests in 5.11s, 1.86GB read
Requests/sec:  20781.65
Transfer/sec:    372.96MB
[info] CPU 6567.0% | Mem 6.9GiB

=== Best: 20911 req/s (CPU: 6765.1%, Mem: 6.8GiB) ===
[info] saved results/static/4096/fastapi.json
httparena-bench-fastapi
httparena-bench-fastapi

==============================================
=== fastapi / static / 6800c (tool=wrk) ===
==============================================
[info] waiting for server...
[info] server ready

[run 1/3]
Running 5s test @ http://localhost:8080
  64 threads and 6800 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   302.15ms  158.69ms   1.60s    74.67%
    Req/Sec   326.63    141.93     1.03k    81.40%
  96518 requests in 5.12s, 1.68GB read
Requests/sec:  18843.05
Transfer/sec:    336.70MB
[info] CPU 6577.7% | Mem 7.6GiB

[run 2/3]
Running 5s test @ http://localhost:8080
  64 threads and 6800 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   311.88ms  109.25ms 771.69ms   81.51%
    Req/Sec   330.36     76.31     0.91k    82.97%
  104961 requests in 5.17s, 1.83GB read
Requests/sec:  20315.55
Transfer/sec:    363.17MB
[info] CPU 6712.9% | Mem 7.9GiB

[run 3/3]
Running 5s test @ http://localhost:8080
  64 threads and 6800 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   312.95ms  110.57ms 785.71ms   81.23%
    Req/Sec   329.08     76.66     0.95k    82.69%
  102865 requests in 5.10s, 1.80GB read
Requests/sec:  20153.35
Transfer/sec:    360.40MB
[info] CPU 6450.0% | Mem 8.0GiB

=== Best: 20315 req/s (CPU: 6712.9%, Mem: 7.9GiB) ===
[info] saved results/static/6800/fastapi.json
httparena-bench-fastapi
httparena-bench-fastapi

==============================================
=== fastapi / async-db / 1024c (tool=gcannon) ===
==============================================
[info] resetting postgres for a clean per-profile baseline
[info] starting postgres sidecar
httparena-postgres
[info] postgres ready (seeded)
[info] waiting for server...
[info] server ready

[run 1/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     1024 (16/thread)
  Pipeline:  1
  Req/conn:  25
  Templates: 5
  Expected:  200
  Duration:  10s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   12.08ms   2.36ms   4.92ms   267.60ms    1.38s

  763544 requests in 10.00s, 763515 responses
  Throughput: 76.33K req/s
  Bandwidth:  287.35MB/s
  Status codes: 2xx=763515, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 763514 / 763515 responses (100.0%)
  Latency overflow (>5s): 1
  Reconnects: 30281
  Per-template: 152199,152902,153113,153190,152110
  Per-template-ok: 152199,152902,153113,153190,152110
[info] CPU 6248.4% | Mem 4.8GiB

[run 2/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     1024 (16/thread)
  Pipeline:  1
  Req/conn:  25
  Templates: 5
  Expected:  200
  Duration:  10s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   12.12ms   2.40ms   4.89ms   260.80ms    1.39s

  760175 requests in 10.00s, 760143 responses
  Throughput: 76.00K req/s
  Bandwidth:  286.11MB/s
  Status codes: 2xx=760143, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 760143 / 760143 responses (100.0%)
  Latency overflow (>5s): 4
  Reconnects: 30241
  Per-template: 151918,152278,152135,151917,151895
  Per-template-ok: 151918,152278,152135,151917,151895
[info] CPU 6325.2% | Mem 4.8GiB

[run 3/3]
gcannon v0.5.3
  Target:    localhost:8080/
  Threads:   64
  Conns:     1024 (16/thread)
  Pipeline:  1
  Req/conn:  25
  Templates: 5
  Expected:  200
  Duration:  10s


  Thread Stats   Avg      p50      p90      p99    p99.9
    Latency   12.32ms   2.33ms   4.76ms   268.70ms    1.44s

  756903 requests in 10.00s, 756870 responses
  Throughput: 75.67K req/s
  Bandwidth:  285.10MB/s
  Status codes: 2xx=756870, 3xx=0, 4xx=0, 5xx=0
  Latency samples: 756869 / 756870 responses (100.0%)
  Latency overflow (>5s): 4
  Reconnects: 30079
  Per-template: 151191,151272,151480,151804,151122
  Per-template-ok: 151191,151272,151480,151804,151122
[info] CPU 6298.6% | Mem 4.8GiB

=== Best: 76351 req/s (CPU: 6248.4%, Mem: 4.8GiB) ===
[info] input BW: 5.10MB/s (avg template: 70 bytes)
[info] saved results/async-db/1024/fastapi.json
httparena-bench-fastapi
httparena-bench-fastapi
[info] skip: fastapi does not subscribe to crud
[info] skip: fastapi does not subscribe to fortunes
[info] skip: fastapi does not subscribe to baseline-h2
[info] skip: fastapi does not subscribe to static-h2
[info] skip: fastapi does not subscribe to baseline-h2c
[info] skip: fastapi does not subscribe to json-h2c
[info] skip: fastapi does not subscribe to baseline-h3
[info] skip: fastapi does not subscribe to static-h3
[info] skip: fastapi does not subscribe to gateway-64
[info] skip: fastapi does not subscribe to gateway-h3
[info] skip: fastapi does not subscribe to production-stack
[info] skip: fastapi does not subscribe to unary-grpc
[info] skip: fastapi does not subscribe to unary-grpc-tls
[info] skip: fastapi does not subscribe to stream-grpc
[info] skip: fastapi does not subscribe to stream-grpc-tls
[info] skip: fastapi does not subscribe to echo-ws
[info] skip: fastapi does not subscribe to echo-ws-pipeline
[info] skip: fastapi does not subscribe to echo-ws-limited
[info] rebuilding site/data/*.json
[updated] /home/diogo/actions-runner/_work/HttpArena/HttpArena/site/data/frameworks.json
[updated] /home/diogo/actions-runner/_work/HttpArena/HttpArena/site/data/results/fastapi.json - 19 new, 19 total
[updated] /home/diogo/actions-runner/_work/HttpArena/HttpArena/site/data/current.json
[info] done
httparena-postgres
[info] restoring loopback MTU to 65536

[Kaliumhexacyanoferrat]

Okay, wow 😄