ci: concurrent sampling ci test

[kapishupadhyay22]

Describe the changes that are made

Adds an end-to-end GitHub Actions pipeline that exercises the --enable-sampling flag of keploy record and asserts the documented bypass behaviour.

What runs on every PR and every push to main:

1. Build keploy from the current branch.
2. Check out the companion sampling-test sample app from keploy/samples-go.
3. Start keploy record --enable-sampling=$K -c ./sampling-test.
4. Wait for the app on :8080, then run curl.sh to fire $N concurrent GET /work requests, each on its own TCP connection. The handler sleeps HANDLER_DELAY_MS (default 500 ms) so the burst overlaps inside the proxy and the slot semaphore is genuinely contended.
5. Drain captures, send SIGINT to keploy.
6. Inspect ./keploy/test-set-0/tests/ and assert:
   • every curl returned 2xx (the bypass path must never break clients — feedback_no_client_errors), and
   • $K ≤ captured_test_cases < $N (sampling actually gated some traffic, but not all of it).

Defaults: K=5, N=20, handler delay 500 ms. All overridable via workflow_dispatch inputs (sampling_limit, total_requests, samples_go_ref).

Files added:

• .github/workflows/sampling-test.yml — workflow definition; triggers on pull_request, push to main, and workflow_dispatch.
• .github/workflows/test_workflow_scripts/sampling/run-sampling-test.sh — driver script (build sample, record, curl burst, SIGINT, assert).

The companion sample app lives in keploy/samples-go#232. The workflow currently pins samples_go_ref to go-sampling so the two PRs can be merged in either order; flip the default to main once the samples-go PR is in.

Links & References

Closes: #4211

🔗 Related PRs

• feat: app for testing concurrent sampling samples-go#232 — feat: app for testing concurrent sampling (the sample app this workflow drives)

🐞 Related Issues

• [feature]: Add e2e CI coverage for --enable-sampling concurrent-request bypass #4211

📄 Related Documents

• NA

What type of PR is this? (check all applicable)

• 📦 Chore
• 🍕 Feature
• 🐞 Bug Fix
• 📝 Documentation Update
• 🎨 Style
• 🧑‍💻 Code Refactor
• 🔥 Performance Improvements
• ✅ Test
• 🔁 CI
• ⏩ Revert

Added e2e test pipeline?

• 👍 yes — this PR is the e2e pipeline

Added comments for hard-to-understand areas?

• 👍 yes — the driver script's count-and-assert block explains why we filter to get-work-*.yaml only

Added to documentation?

• 📜 README.md
• 📓 Wiki
• 🙅 no documentation needed — the sample's README.md (in samples-go) explains how to reproduce locally

Are there any sample code or steps to test the changes?

• 👍 yes, mentioned below

Reproduce locally:

# In keploy/samples-go (on the go-sampling branch)
cd sampling-test
go build -o sampling-test .

# In another terminal
sudo keploy record --enable-sampling=5 -c ./sampling-test &
TOTAL_REQUESTS=20 bash curl.sh

# Then SIGINT keploy and inspect
ls ./keploy/test-set-0/tests/   # expect 5x get-work-*.yaml + 1x get-health-1.yaml

Self Review done?

• ✅ yes

Any relevant screenshots, recordings or logs?

Latest CI run on this branch — successful assertion (post-fix in commit 843d753):

captured /work test cases: 5
total recorded test cases (incl. /health probe): 6
sampling limit (K): 5
total /work requests (N): 20
✅ sampling assertion passed: 5 <= 5 < 20

curl-side, every client got a 2xx response:

curl summary: total=20 ok=20 non2xx=0 wait_failures=0

🧠 Semantics for PR Title & Branch Name

• PR Title: ci: concurrent sampling ci test
• Branch Name: ci-sampling

---

Additional checklist:

• Have you read the Contributing Guidelines on issues?
• Have you followed the PR Semantics guide for naming this PR?
• Have you followed the Branch Semantics guide for naming your branch?

[Copilot]

Pull request overview

Adds a dedicated GitHub Actions workflow to run an end-to-end concurrency sampling check against the samples-go/sampling-test app, validating that --enable-sampling gates recordings under concurrent load without breaking client responses.

Changes:

• Introduces a new CI workflow (Sampling Test) to build Keploy, check out keploy/samples-go, and execute the sampling E2E test.
• Adds a new bash harness that records with --enable-sampling, generates concurrent traffic, stops the recorder, and asserts captured test counts fall within expected bounds.
• Uploads relevant artifacts (record log, recordings, curl results) on failure for debugging.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
.github/workflows/test_workflow_scripts/sampling/run-sampling-test.sh	Adds the end-to-end sampling test driver script (build app, run record with sampling, fire concurrent traffic, assert capture bounds).
.github/workflows/sampling-test.yml	Adds a new workflow to build Keploy, check out samples-go, run the sampling test script, and upload artifacts on failures.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

🚀 Keploy Performance Test Results

Multi-Run Validation: Tests run 3 times, pipeline fails only if 2+ runs show regression.

Run	P50	P90	P99	RPS	Error Rate	Status
1	2.87ms	3.74ms	5.23ms	100.00	0.00%	✅ PASS
2	2.75ms	3.52ms	5.01ms	100.02	0.00%	✅ PASS
3	2.79ms	3.78ms	5.15ms	100.02	0.00%	✅ PASS

Thresholds: P50 < 5ms, P90 < 15ms, P99 < 70ms, RPS >= 100 (±1% tolerance), Error Rate < 1%

✅ Result: PASSED - Only 0 out of 3 runs failed (threshold: 2)

P50, P90, and P99 percentiles naturally filter out outliers

[github-actions]

🚀 Keploy Performance Test Results

Multi-Run Validation: Tests run 3 times, pipeline fails only if 2+ runs show regression.

Run	P50	P90	P99	RPS	Error Rate	Status
1	2.6ms	3.24ms	4.64ms	100.02	0.00%	✅ PASS
2	2.38ms	3.1ms	4.39ms	100.02	0.00%	✅ PASS
3	2.68ms	3.46ms	4.72ms	100.02	0.00%	✅ PASS

Thresholds: P50 < 5ms, P90 < 15ms, P99 < 70ms, RPS >= 100 (±1% tolerance), Error Rate < 1%

✅ Result: PASSED - Only 0 out of 3 runs failed (threshold: 2)

P50, P90, and P99 percentiles naturally filter out outliers

[github-actions]

🚀 Keploy Performance Test Results

Multi-Run Validation: Tests run 3 times, pipeline fails only if 2+ runs show regression.

Run	P50	P90	P99	RPS	Error Rate	Status
1	2.64ms	3.39ms	4.95ms	100.02	0.00%	✅ PASS
2	2.55ms	3.24ms	4.59ms	100.02	0.00%	✅ PASS
3	2.64ms	3.62ms	4.95ms	100.02	0.00%	✅ PASS

Thresholds: P50 < 5ms, P90 < 15ms, P99 < 70ms, RPS >= 100 (±1% tolerance), Error Rate < 1%

✅ Result: PASSED - Only 0 out of 3 runs failed (threshold: 2)

P50, P90, and P99 percentiles naturally filter out outliers

[github-actions]

🚀 Keploy Performance Test Results

Multi-Run Validation: Tests run 3 times, pipeline fails only if 2+ runs show regression.

Run	P50	P90	P99	RPS	Error Rate	Status
1	2.71ms	3.33ms	4.77ms	100.02	0.00%	✅ PASS
2	2.64ms	3.28ms	4.52ms	100.02	0.00%	✅ PASS
3	2.64ms	3.41ms	4.73ms	100.02	0.00%	✅ PASS

Thresholds: P50 < 5ms, P90 < 15ms, P99 < 70ms, RPS >= 100 (±1% tolerance), Error Rate < 1%

✅ Result: PASSED - Only 0 out of 3 runs failed (threshold: 2)

P50, P90, and P99 percentiles naturally filter out outliers