Cross-platform HTTP load testing CLI: a modern ab/wrk alternative with machine-readable reports for CI/CD
cryload is a fast, single-binary HTTP load testing and benchmarking CLI: drive concurrent traffic against REST APIs, microservices, and static sites, measure requests per second, latency percentiles (p50–p999), status breakdowns, and transfer volume. Use it for stress tests, smoke tests before deploy, capacity checks, and GitHub Actions / pipeline automation via JSON or CSV output.
If you are looking for a hey-like or oha-like tool with extra reporting modes, or an ab / wrk alternative for HTTP scenarios without Lua scripting, cryload is built for that workflow. Implemented in Crystal for a small footprint and predictable performance.
Typical uses: bench Node, Go, Python, Rails, or .NET HTTP services; soak an API gateway or Kubernetes ingress; compare p99 latency after tuning; ship the same macOS, Linux, and Windows CLI to your team.
Rough feature snapshot (tools evolve; check each project’s docs for the latest).
| cryload | ab | hey | oha | wrk | |
|---|---|---|---|---|---|
| Language | Crystal | C | Go | Rust | C |
Concurrent connections (-c) |
✓ | ✓ | ✓ | ✓ | ✓ |
Duration / request count (-n) |
✓ | ✓ (-t / -n) |
✓ | ✓ | ✓ |
| JSON / CSV / quiet output for CI/CD | ✓ | — (text) | JSON | JSON | — (text / Lua) |
CI thresholds via exit codes (--max-p99, --fail-on-error) |
✓ | — | — | — | — |
| Text latency histogram + distribution | ✓ | basic | limited | TUI-focused | basic |
Global RPS cap (--rate) |
✓ | — | per-worker (-q) |
✓ | different model |
| Warmup, proxy, cookies, multi-URL file | ✓ | — | partial | partial | — |
| Follow redirects, custom success HTTP codes | ✓ | — | partial | partial | — |
No keep-alive mode (--disable-keepalive) |
✓ | default (enable with -k) |
✓ | ✓ (HTTP/1.1) | via Lua |
| Body from string / file / stdin | ✓ / ✓ / ✓ | file (-p) |
✓ / ✓ / — | ✓ / ✓ / — | via Lua |
| HTTP/2 | — (HTTP/1.1) | — | ✓ (-h2) |
✓ (--http2) |
— |
| Multi-core load generation | — (single-core) | — | ✓ (-cpus) |
✓ | ✓ (-t) |
| Scriptable load (Lua, etc.) | — | — | — | — | ✓ |
Choose wrk when you need Lua-driven scenarios, multi-core saturation, and maximum tuning on Linux. Choose ab when the classic Apache Bench one-liner is enough—plain-text summaries, GET-heavy checks, and httpd-family packages already on the machine. Choose hey or oha when you need HTTP/2 or want to saturate every core of the load-generating machine. Choose cryload when you want CSV / JSON reporting, CI threshold exit codes, rate limits, redirect handling, and histogram-style summaries in one cross-platform binary.
- High-throughput HTTP load testing with a lightweight CLI experience
- Concurrent benchmarking with configurable connection count
- Request count mode (
-n) and duration mode (-d) support - Flexible request customization (method, headers, body from string/file/stdin, auth, user-agent, host header, timeout, TLS, keep-alive control)
- JSON output mode for CI/CD and automation workflows
- Richer latency percentiles plus response/error breakdowns
- Optional global request rate limiting with
--rate - Warmup phase, HTTP(S) proxy, session cookies, multi-URL targets, and cache-busting random paths
- Live progress updates on stderr during text output (disable with
--no-progress)
Downloads the matching asset from Releases, verifies SHA256, and installs to ~/.local/bin (or %USERPROFILE%\.local\bin on Windows).
Linux / macOS (needs curl or wget, and sha256sum or shasum):
curl -sSfL https://raw.githubusercontent.com/sdogruyol/cryload/master/scripts/install.sh | sh -sInstall a specific version:
VERSION=v5.0.0 curl -sSfL https://raw.githubusercontent.com/sdogruyol/cryload/master/scripts/install.sh | sh -sWindows (PowerShell):
iwr -useb https://raw.githubusercontent.com/sdogruyol/cryload/master/scripts/install.ps1 | iexDownload the matching asset from the Releases page (cryload-linux, cryload-linux-arm64, cryload-macos, or cryload-windows.exe).
Linux:
chmod +x cryload-linux
./cryload-linux --helpmacOS: Same as Linux, using the cryload-macos binary.
Windows (PowerShell or Command Prompt, from the directory containing the file):
.\cryload-windows.exe --helpRequires Crystal 1.19.0 or later.
git clone https://github.com/sdogruyol/cryload.git && cd cryload
shards build --releaseThe binary will be at bin/cryload.
Run your first benchmark in seconds:
bin/cryload http://localhost:3000 -n 10000 -c 100cryload <url> [options]Options:
| Option | Description |
|---|---|
-n, --numbers |
Number of requests to make |
-d, --duration |
Duration of test in seconds |
-c, --connections |
Concurrent connections (default: 10) |
-m, --method |
HTTP method (default: GET) |
-b, --body |
HTTP request body |
--body-file |
Read HTTP request body from file |
--body-stdin |
Read HTTP request body from standard input |
-H, --header |
HTTP header, repeatable (-H "Key: Value") |
--user-agent |
Set the User-Agent header |
--host-header |
Override the Host header |
-a, --basic-auth |
HTTP Basic auth in the form user:password |
--timeout |
Client connect/read timeout in seconds |
-q, --rate |
Total request rate limit in requests/sec |
-L, --follow-redirects |
Follow HTTP redirects up to 5 hops |
--disable-keepalive |
Open a new connection for every request (sends Connection: close) |
--output-format |
Output format: text, json, csv, quiet |
--success-status |
Treat specific status codes/ranges as successful |
--insecure |
Accept invalid TLS certificates for HTTPS |
--json |
Print final result as JSON |
--fail-on-error |
Exit with code 1 when any HTTP or transport error occurs |
--fail-on-transport-error |
Exit with code 1 when any transport error occurs |
--max-fail-rate |
Exit with code 1 when failure rate exceeds PERCENT |
--max-p99 |
Exit with code 1 when p99 latency exceeds MS milliseconds |
--warmup |
Warm up before the timed benchmark (seconds) |
--proxy |
HTTP(S) proxy (http://host:port or http://user:pass@host:port) |
--no-progress |
Disable live progress on stderr |
--progress |
Show live progress on stderr during the run (default) |
--cookie |
Cookie value, repeatable (name=value) |
--urls-file |
Load target URLs from file (one http(s) URL per line) |
--random-path |
Append a random path segment to each request URL |
-V, --version |
Print version |
-h, --help |
Show help |
Examples:
10,000 requests to localhost
cryload http://localhost:9292 -n 1000010 seconds with 100 connections
cryload http://localhost:3000 -d 10 -c 100Simple POST request
cryload http://localhost:3000/api/login -n 1000 -m POSTPOST with plain text body
cryload http://localhost:3000/api/echo -n 500 -m POST -H "Content-Type: text/plain" -b "hello"POST with JSON body
cryload http://localhost:3000/api -n 500 -m POST -H "Content-Type: application/json" -b '{"name":"cry"}' --timeout 5POST JSON body from file
cryload http://localhost:3000/api -n 500 -m POST -H "Content-Type: application/json" --body-file payload.jsonPOST body from stdin (pipe-friendly)
jq -c '.payload' fixture.json | cryload http://localhost:3000/api -n 500 -m POST -H "Content-Type: application/json" --body-stdinPOST with multiple headers
cryload http://localhost:3000/api -n 300 -m POST -H "Authorization: Bearer token123" -H "X-Request-ID: benchmark-1" -b '{"ok":true}'Basic auth request
cryload http://localhost:3000/private -n 300 --basic-auth username:passwordCustom User-Agent and Host header
cryload http://127.0.0.1:3000 -n 300 --user-agent cryload-test/1.0 --host-header api.internalDuration mode + timeout
cryload http://localhost:3000/api -d 15 -c 50 --timeout 3Rate-limited run at 100 requests/sec total
cryload http://localhost:3000/api -n 1000 -c 50 --rate 100Warmup before the timed benchmark
cryload http://localhost:3000 -d 30 -c 50 --warmup 5HTTP(S) proxy
cryload http://localhost:3000 -n 100 --proxy http://127.0.0.1:8080
cryload https://api.example.com -n 100 --proxy http://user:pass@proxy.example.com:8080Session cookies (repeatable)
cryload http://localhost:3000/api -n 100 --cookie session=abc123 --cookie theme=darkMultiple URLs from file (positional URL optional)
cryload --urls-file targets.txt -n 1000 -c 50Cache-busting random path on each request
cryload http://localhost:3000/api -n 500 --random-pathMeasure connection setup cost (no keep-alive)
cryload http://localhost:3000/api -n 1000 --disable-keepaliveFollow redirects
cryload http://localhost:3000/redirect -n 100 -LTreat redirects as success without following them
cryload http://localhost:3000/redirect -n 100 --success-status 200-299,302HTTPS with self-signed cert (skip TLS verification)
cryload https://localhost:8443 -n 1000 --insecureJSON output for automation/CI
cryload http://localhost:3000/api -n 1000 --jsonCSV output for scripts
cryload http://localhost:3000/api -n 1000 --output-format csvQuiet mode for exit-code-only checks
cryload http://localhost:3000/health -n 10 --output-format quietCI failure thresholds
cryload http://localhost:3000/api -n 1000 --fail-on-error
cryload http://localhost:3000/api -n 1000 --max-fail-rate 5
cryload http://localhost:3000/api -n 1000 --max-p99 200Print version
cryload --versionExample output (sample run: -d 10 -c 10 against a local server with a 13-byte body):
Preparing to make it CRY for 10 seconds with 10 connections!
Running load test @ http://127.0.0.1:3000
Mode: duration (10s)
Connections: 10
Keep-alive: enabled
Rate limit: unlimited
Warmup: none
Success statuses: 200-299
Summary
Total requests: 161551
Total time: 10.0s
Requests/sec: 16155.06
Responses: 161551
Transport errors: 0 (0.0%)
Fastest: 0.1 ms
Slowest: 2043.12 ms
Status
Successful: 161551 (100.0%)
Failed: 0 (0.0%)
Success statuses: 200-299
Transfer
Total data: 2.0 MiB
Size/request: 13.0 B
Transfer/sec: 205.09 KiB/s
Latency (ms)
avg: 0.62 min: 0.1 stdev: 16.71 max: 2043.1
Latency Percentiles (ms)
p50: 0.3 p90: 0.5 p95: 0.6
p99: 0.9 p999: 1.4
Latency Histogram (ms)
185.8 ms [161512] |■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■
371.6 ms [0] |
557.3 ms [0] |
743.0 ms [0] |
928.8 ms [0] |
1114.5 ms [35] |■
1300.2 ms [3] |■
1485.9 ms [0] |
1671.7 ms [0] |
1857.4 ms [0] |
2043.1 ms [1] |■
Latency Distribution (ms)
10.0% in 0.1
25.0% in 0.2
50.0% in 0.3
75.0% in 0.4
90.0% in 0.5
95.0% in 0.6
99.0% in 0.9
99.9% in 1.4
Status Code Distribution
[200] 161551 responses (100.0%)
cryload is written in Crystal, combining Ruby-like developer ergonomics with compiled-language speed.
Use --json or --output-format csv for scripts, dashboards, and GitHub Actions jobs: parse a single structured summary instead of scraping text. Combine with -n for fixed request counts so pipelines stay deterministic.
| Code | When |
|---|---|
0 |
Run completed and all configured thresholds passed |
1 |
Validation/usage error, or a CI threshold failed |
Default failure behavior: exit 1 only when every request is a transport error (no HTTP responses received).
Optional CI thresholds (combine as needed; any violation exits 1):
| Flag | Effect |
|---|---|
--fail-on-error |
Fail on any HTTP failure or transport error |
--fail-on-transport-error |
Fail on any transport error, even if some requests succeed |
--max-fail-rate 5 |
Fail when (HTTP failures + transport errors) / total requests exceeds 5% |
--max-p99 200 |
Fail when p99 latency exceeds 200 ms |
Examples:
# Smoke test: any 4xx/5xx or connection error fails the job
cryload http://localhost:3000/health -n 50 --fail-on-error --output-format quiet
# SLA gate: p99 must stay under 200 ms, failure rate under 1%
cryload http://localhost:3000/api -n 1000 --max-p99 200 --max-fail-rate 1 --output-format quiet--json / --output-format json emits structured JSON with these sections:
| Section | Fields |
|---|---|
summary |
requests, responses, transport_errors, elapsed_seconds, requests_per_second, failure_rate_percent |
transfer |
total_bytes, size_per_request_bytes, bytes_per_second (response body only) |
latency_ms |
avg, min, max, stdev, p10, p25, p50, p75, p90, p95, p99, p999 |
latency_histogram[] |
start_ms, end_ms, count, percent |
status |
success_statuses, successful_count, successful_percent, failed_count, failed_percent, transport_error_percent, codes[], transport_errors[] |
Also includes top-level url and duration_mode. Full field reference: docs/json-output.md.
- name: Install cryload
run: curl -sSfL https://raw.githubusercontent.com/sdogruyol/cryload/master/scripts/install.sh | sh -s
- name: Verify binary
run: cryload --version
- name: Smoke test API
run: |
cryload http://localhost:3000/health -n 100 --fail-on-error --output-format quiet
- name: Latency SLA check
run: |
cryload http://localhost:3000/api -n 500 --max-p99 250 --json > result.json
jq -e '.summary.failure_rate_percent <= 1' result.json
jq -e '.latency_ms.p99 <= 250' result.jsonNote: --output-format quiet suppresses report output; combine with threshold flags when you only need pass/fail from the exit code.
- Fork the repo
- Create your feature branch (
git checkout -b my-feature) - Commit your changes (
git commit -am 'Add feature') - Push to the branch (
git push origin my-feature) - Open a Pull Request
MIT
