Repository navigation
Add fault injection: simulated errors, timeouts and latency - #50
Merged
Merged
Conversation
Make the emulated bank misbehave on purpose so apps can test error handling, retries and timeouts. Applies to /za and the token endpoint, never to the sandbox's own controls. - Per request: X-Sandbox-Fault: 429|500|502|503|504 (JSON errorMessage, Retry-After on 429), timeout or timeout:<ms> (hold, then drop the connection), reset (drop now); X-Sandbox-Latency: <ms>. Unknown values get a 400. - At random: GET/PUT/DELETE /faults with errorRate + errorStatus, timeoutRate, latencyMs and an optional pathPrefix; saved in the Setting table across restarts and broadcast to dashboards. - Faulted responses carry X-Sandbox-Fault-Injected and show in the live server log (new logRequest helper, also used by formatErrorResponse). - The dashboard tags its requests X-Sandbox-Client: dashboard and is exempt from random faults; it gets a Fault injection panel and a banner with Turn off while faults are on. - OpenAPI: the /faults endpoints, and the headers in the API description. Tests: 20 API (each status, reset, timeout, latency, token endpoint, validation, rates, path prefix, dashboard exemption, persistence), 2 unit, 1 e2e. 203 API, 41 unit and 17 e2e passing. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
- The CodeQL Autofix commit (5c30ac6) renamed the parsed X-Sandbox-Latency value to requestedLatencyMs but dropped the assignment to latencyMs, so the header was validated and then ignored. That is what failed the latency test in CI. Restored the assignment (keeping the validated, bounded name). - CodeQL (js/resource-exhaustion): the timeout fault's setTimeout is now bounded where the timer is created, not only when parsing the header, and at most 25 requests are held open at once; more timeout faults drop the connection straight away (logged), so a client can't tie up the server with held connections. - Fault tests time with performance.now() (monotonic) instead of Date.now(). - New test: 26 concurrent timeout faults, one dropped at once, 25 held. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Make the emulated bank misbehave on purpose, to test an app's error handling, retries and timeouts, which the official sandbox can't do. Faults apply to
/za/...and the token endpoint, never to the sandbox's own controls, so they can always be turned off.Per request (deterministic)
X-Sandbox-Fault: 429/500/502/503/504{ errorMessage: "Simulated fault: …" }; 429 addsRetry-After: 1X-Sandbox-Fault: timeout/timeout:2000X-Sandbox-Fault: resetX-Sandbox-Latency: 1500Unknown values get a 400 that explains the allowed ones.
At random
GET/PUT/DELETE /faultsseterrorRatewitherrorStatus,timeoutRate,latencyMs, and an optionalpathPrefix(e.g. only/za/v1/cards).Visibility
X-Sandbox-Fault-Injected, and every fault, including dropped connections, appears in the live server log.Dashboard
X-Sandbox-Client: dashboard, so random faults don't break it while your app sees them. Explicit fault headers still apply.Docs
The
/faultsendpoints are in OpenAPI, and the headers are described in the API description, so they're visible in Swagger. The README has a new section.Test plan
test/faults.spec.tscovers each status, reset and timeout really dropping the connection, latency adding time, the token endpoint, header validation, 100% and ~50% rates, the path prefix, the dashboard exemption, persistence, and that/faultsitself is never faulted.describeFaults), lint, typecheck, build; 17 Playwright tests. New: turn faults on in the panel, an outside request gets 503 while the dashboard keeps working, Turn off from the banner.openapi:check(64 routes), knip🤖 Generated with Claude Code