Files
ui/_docs/02_document/tests/resilience-tests.md
Oleksandr Bezdieniezhnykh f7dd6c98d8
ci/woodpecker/push/build-arm Pipeline failed
[AZ-501] [AZ-502] Cycle 2 Step 14 security audit + inline fixes
Security audit (5 phases) → reports under _docs/05_security/.

AZ-501 (F-SAST-1, HIGH): Externalize hardcoded Google Geocode key
from mission-planner/src/config.ts to VITE_GOOGLE_GEOCODE_KEY via
new GeocodeService.ts; fail-soft warn when unset; STC-SEC1D static
deny-list gate; +5 unit tests in tests/mission_planner_geocode.test.ts.

AZ-502 (F-DEP-1, HIGH): Force vite>=6.4.2 and postcss>=8.5.10 via
package.json overrides in both roots; clean reinstall clears all
bun audit advisories.

Test-spec sync (Step 12) + Update Docs (Step 13) deltas: AC-43, AC-44,
NFT-SEC-09b, FT-P-61, FT-N-17, ripple log, batch_12 report.

Pending user actions: revoke Google + OWM keys (AC-6 / AZ-499 AC-7).

229 PASS / 13 SKIP / 0 FAIL on static + fast suites.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-05-12 05:31:11 +03:00

11 KiB

Resilience Tests

Failure / recovery scenarios at the SPA's observable boundary: bearer expiry, refresh cookie loss, upstream 5xx, network partition, oversized uploads, SSE drop. Every fault injection is at the network / browser layer; the UI is observed for graceful behavior and recovery.

NFT-RES-01: 401 → refresh → retry recovery is transparent

Summary: An authenticated request that returns 401 mid-session is refreshed and retried without unmounting the routed view. Traces to: AC-01, AC-23

Preconditions:

  • Active session on /flights.

Fault injection:

  • Force a 401 on the next outbound authenticated request.

Steps:

Step Action Expected Behavior
1 Stub the next request to return 401 once first response 401
2 Observe the SPA's reaction POST /api/admin/auth/refresh with credentials:'include'; on 200, original request retried with new bearer
3 Inspect <ProtectedRoute> children not unmounted; re-render delta ≤ 1

Pass criteria: row 03 sequence; row 11 re-render bound (≤ 1); row 12 refresh count == 1 — all hold simultaneously. Expected result source: results_report.md rows 03, 11, 12.


NFT-RES-02: SSE bearer-rotation — both streams reconnect within 5 s

Summary: A bearer rotation during open SSE streams (live-GPS + annotation-status) tears them down and reopens them with the new token. Traces to: AC-24

Preconditions:

  • Two EventSources open.

Fault injection:

  • Trigger a server-driven rotation of the bearer (force a refresh).

Steps:

Step Action Expected Behavior
1 Rotate bearer new token in memory
2 Observe each EventSource closes, then reopens with new ?token=
3 Measure max reconnect time ≤ 5 000 ms

Pass criteria: both streams close+open exactly once; max reconnect ≤ 5 000 ms (results_report.md row 13). Status: quarantined until SSE reconnect-on-rotation ships (Step 8 hardening). Expected result source: results_report.md row 13.


NFT-RES-03: Network offline at boot — error state, no offline mode

Summary: With network disabled, app boot results in a user-visible error state — NOT a service worker-served cached UI. Traces to: AC-N3

Preconditions:

  • Browser network disabled (or all /api/* stubs respond with offline error).

Fault injection:

  • All outbound requests fail with net::ERR_INTERNET_DISCONNECTED (or equivalent).

Steps:

Step Action Expected Behavior
1 Load the SPA static assets served by nginx OK; API calls fail
2 Observe DOM login or general error surface present
3 Inspect navigator.serviceWorker.controller null

Pass criteria: row 93 — error/login-failed state present; no service worker controller. Expected result source: results_report.md row 93.


NFT-RES-04: ProtectedRoute loading timeout fallback after 10 s

Summary: The <ProtectedRoute> spinner has a bounded loading window; a stalled auth bootstrap surfaces a retry CTA / error. Traces to: AC-17

Preconditions:

  • Bootstrap refresh stubbed to never resolve.

Fault injection:

  • POST /api/admin/auth/refresh hangs (no response).

Steps:

Step Action Expected Behavior
1 Mount <ProtectedRoute> spinner rendered
2 Advance fake time to 10 s timeout fires
3 Inspect DOM retry CTA or error message present; spinner unmounted

Pass criteria: row 59 — fallback present, spinner absent. Status: quarantined until timeout fix lands (Step 4). Expected result source: results_report.md row 59.


NFT-RES-05: Settings save with upstream 500 — UI state recovers

Summary: A 500 on settings save surfaces an error and resets the saving flag. Traces to: AC-27

Preconditions:

  • Form filled in valid state on /settings.

Fault injection:

  • Upstream PUT /api/annotations/settings/system returns 500 after T ms.

Steps:

Step Action Expected Behavior
1 Click Save PUT issued
2 Stub responds 500 within 2 s failure
3 Inspect within 2 s toast / inline error; saving === false; no route navigation

Pass criteria: row 68. Status: quarantined until Step 4 try/finally fix. Expected result source: results_report.md row 68.


NFT-RES-06: Settings save with network drop — try/finally state reset

Summary: When the underlying fetch throws (network drop), saving resets and the user sees an error. Traces to: AC-27

Preconditions:

  • Form filled in valid state.

Fault injection:

  • Network drop mid-PUT (fetch rejects).

Steps:

Step Action Expected Behavior
1 Click Save PUT issued
2 Stub throws rejection delivered
3 Inspect saving === false; error surfaced

Pass criteria: row 69. Status: quarantined. Expected result source: results_report.md row 69.


NFT-RES-07: nginx 413 on oversized upload surfaces user-visible error

Summary: An upload that exceeds client_max_body_size 500M returns 413; the UI presents a user-facing message (no silent failure, no alert()). Traces to: AC-10

Preconditions:

  • Authenticated; <MediaList> open.

Fault injection:

  • Drop a 501 MB synthetic file.

Steps:

Step Action Expected Behavior
1 Upload 501 MB upload starts
2 nginx rejects 413 delivered
3 Inspect UI error containing the i18n "file too large" string; no alert() invoked

Pass criteria: row 39. Expected result source: results_report.md row 39.


Summary: When the refresh cookie is gone (or expired) and a 401 occurs, the SPA redirects the user to /login rather than silently looping refresh. Traces to: AC-01, AC-22

Preconditions:

  • Cookie cleared from the browser jar.

Fault injection:

  • 401 on any authenticated call; subsequent POST /api/admin/auth/refresh returns 401.

Steps:

Step Action Expected Behavior
1 Issue authenticated call 401
2 Refresh attempted 401 (no cookie)
3 Observe routing redirect to /login

Pass criteria: final URL /login; no infinite refresh loop (single refresh attempt). Derived from AC-01 + AC-22; no specific results_report row binds the loop bound — Phase 3 flags this and the loop bound is added to row 03 if accepted.


NFT-RES-09: Annotation download tainted-canvas fallback

Summary: When <canvas>.toBlob() raises a tainted-canvas exception (cross-origin video frame), the user sees an error rather than a silent no-op. Traces to: NFR (04_verification_log.md finding on handleDownload tainted-canvas risk)

Preconditions:

  • An annotation is loaded from a video sourced with CORS that taints the canvas.

Fault injection:

  • Cross-origin video source taints the canvas; toBlob throws.

Steps:

Step Action Expected Behavior
1 Click Download export attempted
2 toBlob throws error handled
3 Inspect UI user-visible error; no alert(); no silent swallow

Pass criteria: row 96 — error surfaced; no silent swallow; no fabricated blob; no alert(). Expected result source: results_report.md row 96.


NFT-RES-10: SSE server disconnect — UI surfaces a connection-lost indicator

Summary: When the suite server closes a live-GPS or status-events SSE without rotation, the UI does NOT show stale data and DOES indicate the connection lost. Traces to: AC-08, AC-09, AC-24

Preconditions:

  • One SSE stream open.

Fault injection:

  • Server force-closes the stream (no rotation).

Steps:

Step Action Expected Behavior
1 Drop the stream from the server error fires on EventSource
2 Observe DOM a connection-lost indicator is rendered (or stale-data badge)
3 Observe reconnect behavior EventSource auto-retries per browser default; if the SPA re-creates it, exactly one new instance

Pass criteria: row 97 — connection-lost indicator OR reconnect attempt within 10 s; stale data NOT rendered as live; reconnect attempts ≤ 1 in the 10 s window. Expected result source: results_report.md row 97.


NFT-RES-11: Tile endpoint 401/503 does NOT crash the map

Summary: When the satellite-provider /tiles/{z}/{x}/{y} endpoint returns 401 (cookie-auth failure) or 503 (Google Maps upstream down), the SPA renders a broken-tile placeholder for the failing tile(s) and the rest of the application keeps working. No React error boundary fires; no full-page crash. Traces to: AC-41 (AZ-498 NFR-Reliability)

Preconditions:

  • <FlightMap> mounted with a valid VITE_SATELLITE_TILE_URL.
  • Tile endpoint configured to return 401 (auth failure) OR 503 (upstream provider down) for one or more tile coordinates.

Fault injection:

  • (auth-failure variant) Strip / invalidate the satellite-provider auth cookie before the SPA attempts a tile fetch; tile endpoint responds 401.
  • (upstream-down variant) Configure the test stub to return 503 for GET /tiles/{z}/{x}/{y}.

Steps:

Step Action Expected Behavior
1 Mount <FlightMap>; trigger a tile load that fails per the fault Leaflet emits a tileerror event for the affected coordinate
2 Observe the rendered map broken-tile placeholder shown in the failing cell; surrounding tiles continue rendering normally
3 Observe the rest of the SPA (header, side panels, navigation) remains interactive; no React error boundary fires; no console error of category Uncaught
4 Observe a recovery path (auth restored OR upstream back) next pan/zoom successfully fetches the tile; the placeholder is replaced with the imagery

Pass criteria:

  • 401 response on a tile request MUST NOT crash the map; broken-tile placeholder rendered in the failing cell, rest of SPA interactive.
  • 503 response treated identically to 404/transient failure (fault budget — recovery path works after the upstream returns).
  • No new uncaught error in the console attributable to the failed tile.

Expected result source: AZ-498 NFR-Reliability (no results_report.md row needed — observable through DOM state and console). Note on follow-up: AZ-498 risk #5 flags an optional tileerror listener on <MapContainer> that surfaces a structured warning + an optional inline banner ("Imagery unavailable; please re-sign-in"). If/when that lands, this scenario gains a Step 5 asserting the banner appears within 2 s of the first tile error.