SS-53 · Test suites, device farm & quality gates

Continuous TestGrid & device lab

Run a suite; produce signed release evidence.

27operations
5sub-capabilities
4endpoint groups
4services it composes

What it does

Run a suite; produce signed release evidence.

The test and device lab behind a release: suites and test runs, a real-device farm with reservations and sessions, performance budgets, accessibility audits, flaky-test quarantine and human-in-the-loop sessions. A quality-gate evaluation turns run results into a pass/fail verdict, and release evidence is retrievable per release for audit.

In-mesh base http://ss-53-testgrid.shared-services.svc.cluster.local:8053 · auth is an SS-01 bearer token + X-Tenant-Id on every /api/v1 route.

Sub-capabilities

The real building blocks.

Suites & test runs

Define reusable suites and open test runs against them; post results, retry, and attach evidence to a run.

POST /api/v1/test-runs

Real-device farm

Reserve real devices and open device sessions, so a run executes on hardware, not just emulators.

POST /api/v1/device-farm/reservations

Quality-gate evaluation

Evaluate accumulated results against a gate to produce the pass/fail verdict a release is gated on.

POST /api/v1/quality-gates/evaluate

Performance & accessibility

Declare performance budgets and run against them; run accessibility audits with retrievable findings.

POST /api/v1/performance-runs

Flaky quarantine & HIL

Detect flaky tests and quarantine them so they stop blocking merges; open human-in-the-loop sessions for judgment calls.

POST /api/v1/flaky-tests/{id}/quarantine

Worked example

Define a suite → run on real devices → gate the release

A suite runs on the device farm, accumulates results, and a quality-gate evaluation produces the verdict that gates the release — with signed evidence retained.

  1. 1
    Define the suitePOST/api/v1/suites

    A reusable suite of cases.

  2. 2
    Reserve devicesPOST/api/v1/device-farm/reservations

    Real hardware, not just emulators.

  3. 3
    Open the runPOST/api/v1/test-runs

    Run the suite; post results as they land.

  4. 4
    Evaluate the gatePOST/api/v1/quality-gates/evaluate

    Turn accumulated results into a pass/fail verdict.

  5. 5
    Pull release evidenceGET/api/v1/evidence/{release}

    Signed, retrievable evidence for audit.

Endpoint catalog

27 documented endpoints of 27 operations.

Real method + path from services/testgrid/openapi.json. The full contract, including ops endpoints, is the 27-operation surface.

Suites & test runs

  • GET/api/v1/suitesList test suites
  • POST/api/v1/suitesDefine a test suite
  • GET/api/v1/suites/{id}Fetch one suite
  • GET/api/v1/test-runsList test runs
  • POST/api/v1/test-runsOpen a test run against a suite
  • GET/api/v1/test-runs/{id}Fetch one test run
  • POST/api/v1/test-runs/{id}/resultsPost results into a run
  • POST/api/v1/test-runs/{id}/retryRetry a run
  • POST/api/v1/test-runs/{id}/evidenceAttach evidence to a run

Device farm & HIL

  • GET/api/v1/device-farm/reservationsList device reservations
  • POST/api/v1/device-farm/reservationsReserve real devices
  • POST/api/v1/device-sessionsOpen a device session
  • GET/api/v1/hil-sessionsList human-in-the-loop sessions
  • POST/api/v1/hil-sessionsOpen a human-in-the-loop session

Quality, performance & a11y

  • POST/api/v1/quality-gates/evaluateEvaluate results into a pass/fail verdict
  • POST/api/v1/performance-budgetsDeclare a performance budget
  • GET/api/v1/performance-runsList performance runs
  • POST/api/v1/performance-runsRun against a performance budget
  • GET/api/v1/accessibility-auditsList accessibility audits
  • POST/api/v1/accessibility-auditsRun an accessibility audit
  • GET/api/v1/accessibility-audits/{id}Fetch one accessibility audit

Flaky tests & evidence

  • GET/api/v1/flaky-testsList detected flaky tests
  • POST/api/v1/flaky-testsRecord a flaky-test signal
  • POST/api/v1/flaky-tests/{id}/quarantineQuarantine a flaky test
  • GET/api/v1/evidence/{release}Retrieve signed release evidence
  • GET/api/v1/usageTenant usage ledger + SS-10 delivery status
  • GET/api/v1/infoService info + capability envelope

Architecture placement

Where it sits in the platform.

Continuous TestGrid & device lab does not reimplement the platform — it composes it. These are the services it leans on and the seams between them.

SS-47 SkyForge

A green TestGrid gate is a natural pre-condition for a SkyForge auto-deploy on a release branch.

SS-31 Factory

TestGrid results feed the gate-stack verdicts the Factory merge decision requires green.

SS-34 GRC

Signed release evidence per release is exported as compliance evidence.

SS-10 Billing

Device-farm and run usage is metered into the tenant usage ledger.