fiber, fiber-tuned, fiber-websocket, go-fasthttp: Fiber 3.6.0 and fasthttp 1.75.0

[ReneWerner87]

Description

Dependency bump for the three Fiber entries and for go-fasthttp.

entry	change
fiber, fiber-tuned, fiber-websocket	Fiber v3.5.0 -> v3.6.0
fiber, fiber-websocket	fasthttp v1.73.0 -> v1.75.0, the version Fiber 3.6.0 requires
fiber-tuned	fasthttp master pin (c96f600) -> v1.75.0
go-fasthttp	fasthttp v1.69.0 -> v1.75.0

Go 1.26. Fiber 3.6.0 declares go 1.26.0, so fiber and fiber-tuned move their go
directive up from 1.25.0. The Dockerfiles already build on golang:1.26-alpine and are
unchanged.

fiber-tuned loses its fourth difference. It pinned fasthttp master for the switch
from andybalholm/brotli to go-brrr. fasthttp v1.75.0 carries that switch and Fiber 3.6.0
requires v1.75.0, so the standard entry now gets go-brrr through the release and the pin is
gone. Both READMEs, the header comment in fiber-tuned/main.go and its meta.json
description now name three tuned things instead of four. The measurement tables in
fiber-tuned/README.md stay and are labelled as taken on Fiber 3.5.0; they were not
repeated.

Compress is mounted on /json only. Fiber 3.6.0's compress middleware honours q-values,
so it now acts on the br;q=1, gzip;q=0.8 the static profiles send, which 3.5.0 passed to a
matcher that compares whole tokens. Left on /static, it would encode per request whatever
has no twin on disk and a compressible type: the two woff2 fonts, which are compressed already
(bold.woff2, 22,528 bytes, comes out as 22,518 at level 4 and as 22,532 at level 0). So
fiber and fiber-tuned take /static out of the mount, and staticFile sets
Vary: Accept-Encoding itself on every answer, which the middleware did before. Under the
profiles' header every one of the 20 files answers as it does on main, headers included.

go-fasthttp. The fasthttp bump only, no code change. The entry compresses with
compress/gzip itself, so the brotli library switch does not reach it. erikdubbelboer is
added to maintainers.

Checks. gofmt, go vet, go mod verify and go mod tidy -diff are clean for all
four entries and the Docker images build. The images from main and from this branch ran
side by side and were compared on status, headers and decoded body: 206 requests each for
fiber and fiber-tuned (JSON, gzip and brotli, TLS, async-db, crud, traversal paths, and
all 20 static files over HTTP and TLS under four Accept-Encoding variants), 37 for
go-fasthttp. What differs:

• the brotli json-comp body of fiber, 1490 -> 1489 bytes, which is go-brrr at level 4
• /json requested with q-values is now brotli-encoded, which is the 3.6.0 negotiation
• the two woff2 fonts requested with a bare gzip or br token are no longer encoded by
  the middleware and go out as they are on disk

scripts/validate-ws.py passes 7/7 against fiber-websocket. scripts/validate.sh was not
run locally (macOS, no host networking), and no benchmark results are included.

---

PR Commands - comment on this PR to trigger (requires collaborator approval):

Command	Description
/benchmark -f <framework>	Run every test the framework subscribes to
/benchmark -f <framework> -t <test>	Run one test only
/benchmark -f <framework> --save	Run and save results (updates the leaderboard on merge)
/benchmark -f <framework> -t <test> --save	Run one test and save results
/benchmark -f <framework> --compare <other>	Measure the deltas against another framework instead of this one
/benchmark-multiple -f <fw1>,<fw2>,...	Benchmark several frameworks in one run - takes -t and --save too; saved results land in a single commit
/benchmark-multiple --save	No -f needed: benchmark and save every framework the PR touches
/benchmark-test -t <test>	Benchmark all enabled frameworks subscribed to <test> and save the results

For /benchmark, always specify -f <framework>; the flags combine in any order. Results come back as a comment with a per-profile table of RPS, p99, CPU and memory - one table per framework on multi runs. A new benchmark comment while a run is in flight queues behind it (one deep) instead of cancelling it. For multi-framework PRs (dependency bumps, same-language refactors) prefer /benchmark-multiple, which runs everything in a single job and commits all saved results together, so no run overwrites another. --compare works on single-framework runs only.

What the deltas are measured against. By default, this framework's own results published on main - answering "did this change help?". When you are tuning a variant or a successor entry, --compare re-bases them on another entry instead:

/benchmark -f genhttp --compare genhttp-kestrel

The reply states which baseline it used, and profiles the other framework does not run show n/a rather than a delta.

---

Run benchmarks locally

You can validate and benchmark your framework locally with the lite script - no CPU pinning, fixed connection counts, all load generators run in Docker.

./scripts/validate.sh <framework>
./scripts/benchmark-lite.sh <framework> baseline
./scripts/benchmark-lite.sh --load-threads 4 <framework>

Requirements: Docker Engine on Linux. Load generators (gcannon, h2load, h2load-h3, wrk) are built as self-contained Docker images on first run.

[github-actions]

👋 Heads up! This PR modifies the following frameworks:

• fiber → @gaby @sixcolors @efectn @Fenny
• fiber-tuned → @gaby @sixcolors @efectn @Fenny
• fiber → @gaby @sixcolors @efectn @Fenny

[MDA2AV]

/benchmark-multiple --save

[github-actions]

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

[github-actions]

Benchmark Results

Frameworks: 4 | Test: all tests

✅ fiber

Test	Conn	RPS	Rate	CPU/req	p99	p99.9	CPU	Mem	Δ RPS	Δ Mem
baseline	4096	2,623,678	—	—	—	—	6367.5%	348MiB	+2.1%	+3.9%
pipelined	4096	18,543,424	—	—	—	—	6689.5%	357MiB	+7.8%	+5.6%
limited-conn	4096	2,062,489	—	—	—	—	6444.1%	892MiB	+0.6%	+15.8%
json-comp	4096	290,414	—	—	—	—	6238.2%	804MiB	+63.1%	+6.1%
json-comp	16384	288,805	—	—	—	—	6558.8%	1.6GiB	+61.0%	+6.7%
json-tls	4096	902,093	—	—	—	—	6311.6%	722MiB	+6.4%	+10.6%
8gbit	512	49,441	0.9888	67.23us	137.0us	3389.0us	351.8%	406MiB	+0.5%	+36.2%
static-tls	1024	579,542	—	—	—	—	6489.9%	529MiB	+4.0%	+2.9%
async-db	1024	240,295	—	—	—	—	5403.0%	605MiB	+0.8%	+5.4%
latency-1m	1024	997,737	0.9977	38.31us	246.0us	2469.0us	4332.5%	278MiB	~0%	+2.6%
latency-10k	1024	9,983	0.9983	43.56us	106.0us	229.0us	41.8%	275MiB	~0%	+5.8%
latency-500k-8cpu	1024	377,273	0.7545 ⚠️	21.14us	6785024.0us	8515584.0us	789.0%	61MiB	+0.8%	-11.6%
async	32000	1,519,042	—	—	—	—	4208.2%	1.0GiB	~0%	+2.9%

✅ fiber-tuned

Test	Conn	RPS	Rate	CPU/req	p99	p99.9	CPU	Mem	Δ RPS	Δ Mem
baseline	4096	2,539,295	—	—	—	—	6645.1%	911MiB	+1.7%	+0.3%
pipelined	4096	18,511,670	—	—	—	—	6529.6%	906MiB	+9.5%	+1.2%
limited-conn	4096	2,036,184	—	—	—	—	6218.3%	1.2GiB	+1.6%	+20.0%
json-comp	4096	650,735	—	—	—	—	6322.8%	1.1GiB	-1.3%	+10.0%
json-comp	16384	636,030	—	—	—	—	6274.7%	1.9GiB	-0.9%	+18.7%
json-tls	4096	1,135,135	—	—	—	—	6270.8%	1.0GiB	+7.8%	+6.4%
8gbit	512	49,435	0.9887	66.97us	142.0us	3641.0us	349.5%	865MiB	+0.2%	+3.0%
static-tls	1024	573,390	—	—	—	—	6496.8%	842MiB	+4.9%	+2.2%
async-db	1024	240,544	—	—	—	—	4408.4%	929MiB	-0.7%	+2.3%
latency-1m	1024	998,244	0.9982	37.65us	241.0us	1366.0us	4285.5%	884MiB	~0%	+1.8%
latency-10k	1024	9,982	0.9982	43.15us	103.0us	176.0us	41.1%	873MiB	~0%	+0.2%
latency-500k-8cpu	1024	376,681	0.7534 ⚠️	21.17us	6301696.0us	7800832.0us	791.2%	108MiB	+2.0%	-7.7%
async	32000	1,520,864	—	—	—	—	4197.7%	1.3GiB	+0.2%	+8.3%

✅ fiber-websocket

Test	Conn	RPS	CPU	Mem	Δ RPS	Δ Mem
echo-ws	512	3,011,629	6431.5%	338MiB	-5.4%	+0.3%
echo-ws	4096	3,006,069	6379.6%	449MiB	-6.2%	-0.4%
echo-ws	16384	2,649,499	6629.0%	990MiB	-4.8%	+15.4%
echo-ws-pipeline	512	4,621,699	6274.6%	418MiB	-2.6%	+10.9%
echo-ws-pipeline	4096	4,638,201	6071.3%	572MiB	-2.2%	+25.7%
echo-ws-pipeline	16384	4,627,739	6270.7%	1.0GiB	-2.2%	~0%
echo-ws-limited	512	1,655,701	5898.9%	490MiB	-2.3%	+2.9%
echo-ws-limited	4096	1,910,325	6220.7%	633MiB	-4.6%	+3.8%

✅ go-fasthttp

Test	Conn	RPS	Rate	CPU/req	p99	p99.9	CPU	Mem	Δ RPS	Δ Mem
baseline	4096	1,187,086	—	—	—	—	4003.4%	154MiB	-0.7%	-2.5%
pipelined	4096	17,058,051	—	—	—	—	6733.1%	146MiB	+14.9%	-5.2%
limited-conn	4096	694,242	—	—	—	—	4626.5%	307MiB	-1.8%	-0.6%
json-comp	4096	34,340	—	—	—	—	3075.1%	3.2GiB	-0.5%	~0%
json-comp	16384	27,453	—	—	—	—	4043.6%	5.5GiB	+1.4%	-15.4%
json-tls	4096	509,505	—	—	—	—	4923.6%	445MiB	-3.6%	+4.5%
8gbit	512	49,322	0.9864	66.54us	155.0us	7662.0us	351.4%	112MiB	~0%	+13.1%
static-tls	1024	290,357	—	—	—	—	4356.0%	347MiB	+1.5%	+5.8%
async-db	1024	60,466	—	—	—	—	3781.3%	210MiB	+5.5%	+2.9%
latency-1m	1024	998,359	0.9984	33.25us	1129.0us	6618.0us	3630.6%	88MiB	~0%	+4.8%
latency-10k	1024	9,982	0.9982	32.39us	105.0us	2049.0us	30.8%	75MiB	~0%	-15.7%
latency-500k-8cpu	1024	380,114	0.7602 ⚠️	20.98us	4704256.0us	4798464.0us	788.6%	45MiB	NEW	NEW
async	32000	932,469	—	—	—	—	3915.3%	794MiB	NEW	NEW