LensReading which lens this session carries.
All controls

CTL-018 · Change and release

Production and sandbox estates stay separate

The sandbox holds no production data and the live estate accepts no change that was not rehearsed.

IneffectiveTester is not independent
The tester of record is the control owner. A control tested by the team that operates it is not independently tested. This appears on the duties register as a segregation-of-duties conflict, and it is stated here rather than being hidden behind a pass.
Domain
Change and release
Scope
enterprise
Frequency
monthly
Operation
Manual
Control owner
Platform Engineering Lead · engineering
Tester of record
self
Last tested
Aug 07, 2026
Next due
Sep 06, 2026 · in 19d

What counts as evidence for this control

Written before the test, so a tester cannot choose the evidence that suits the result

Environment register and the promotion log for the period.

Test history

3 tests on record

117 items sampled
TestPeriodSample methodSizePassFailResultTesterTested
CT-018-01Q3 month 2 2026random attribute sample47452Pass with findingsPlatform Engineering Lead (own team)
Owner organization — not independent
Aug 07, 2026
CT-018-02Q3 month 1 2026random attribute sample47461Pass with findingsPlatform Engineering Lead (own team)
Owner organization — not independent
Jul 07, 2026
CT-018-03Q2 2026random attribute sample23230PassPlatform Engineering Lead (own team)
Owner organization — not independent
Jun 06, 2026

Exceptions raised against this control

1 still open of 1

EXC-006Sandbox estate is not a true copy of productionmediumopen · due Sep 06, 2026

The sandbox carries a reduced estate, so a change rehearsed there does not prove the change is safe in the live estate at full volume.

Owner
Platform Engineering Lead · engineering
Remediation
Rehearse against a masked full-volume copy, or state openly that rehearsal covers correctness and not scale.

What this page is, and what it is not

Read this before quoting a result

The test results on this page are a modeled test history for a modeled estate. They demonstrate the shape of independent control testing — a named tester outside the operating team, a stated sample method, a sample size, a pass and fail count, and evidence citations that resolve into the chain — not the audited results of a live enterprise.

12 of 12 evidence citations on this control resolve to a record in the evidence spine. Every citation resolves; each one links through to the sealed record it refers to.

A control shown as untested has no result at all. Coverage on this platform counts controls that were actually tested; it never counts a control as passing because it was designed.