InnerWeb
Skip to main content

Checklist for accepting a safety system update

After a safety system update, the plant should confirm that everything works before people return to normal work. The checklist sets the order: version, test result, alarms, logs and acceptance record, with a comment on why each item matters.

242
tests in the E2E package
2-2.5 h
full end-to-end test run

Before you start the acceptance

Acceptance of an update only makes sense if you know what was changed and which version is now running. Start with the vendor's release notice: the version number, the list of changes and any remarks about risk. In error reports the release version is the short hash of the commit of the deployed image, the same as the image tag in the registry, so if a problem occurs the bug can be tied to a specific change.

Schedule the acceptance outside the hours of greatest traffic. An update on the InnerWeb side does not replace the client's readers or cards, so the acceptance concerns the software and the integrations, not the access control hardware.

Checkpoints and why they matter

The full list is in the checklist box below. Here we discuss the points that are most often skipped.

Automated test result

The E2E package includes 242 tests in 31 files and runs a real Chromium browser against a real backend. The tests replay user actions: logging in, the admin panel, work scheduling, work permits, reports, documentation, maps and analytics. Ask the vendor for the result of the last run and check that the error list is empty or contains only known problems.

The full suite takes about 2 to 2.5 hours, so it will not always be run just before acceptance. If the result comes from a weekend run, make sure it concerns the same version you are accepting.

Errors in the browser and on the server

A test fails automatically when an uncaught JavaScript exception or a server response with status 5xx appears in the browser. During your own acceptance, watch for the same signals: server error messages, blank screens and no response to clicks.

Safety alarms

Work safety alarms are not silenced by a cache failure: if a failure occurs, the alarm still rings and the log records the problem. After the update, check that a test alarm reaches the right people and that the log contains no entries about a cache problem.

Test database and production data

The test database is rebuilt from migrations once per test run, so the result does not depend on data from the previous run. Acceptance on production, however, concerns the plant's data, so separately check that permits, maps and documentation open as they did before the update.

Breakdown of checklist points

Breakdown of checklist pointsPie chart: two points before acceptance, three points on tests and errors, three on alarms and data, two on proof of acceptance.Before acceptance: 2 points (20%)Tests and errors: 3 points (30%)Alarms and data: 3 points (30%)Proof of acceptance: 2 points (20%)10pointsBefore acceptance2 points (20%)Tests and errors3 points (30%)Alarms and data3 points (30%)Proof of acceptance2 points (20%)
Source: Summary of the checklist points on this page

When to consider the acceptance failed

Acceptance fails when any critical point fails: a test alarm does not arrive, an integration does not accept events, or the log shows 5xx responses during ordinary actions. In such a case do not return to normal work on the new version. Instead, report the bug to the vendor with the version number and the steps, and agree with the vendor on rolling back the change until a fix is available.

Auxiliary points, for example the appearance of a screen, may be accepted conditionally if the vendor commits to fixing them in the next release. Record the condition in the protocol with the ticket number.

How to record proof of acceptance

For the information security department the trail matters: who accepted, when, which version and with what result. Record the date, the time, the version number, the person accepting and the list of points marked as passed or failed. For a failed point, add the vendor's ticket number.

A known, not yet fixed bug is marked in the tests as an expected failure with a ticket number, never by skipping the test. Apply the same rule in acceptance: enter a known bug in the register with its number, and do not skip it silently.

Checklist

Video

Frequently asked questions

When can we return to normal work after an update?

When all critical points pass: the version, the tests, the test alarm and the integrations. Points that fail go into the register with a ticket number.

Must acceptance be repeated after every release?

Yes, but its scope can match the scope of the changes. The automated test package reduces the number of points to check by hand.

Does the update affect access control hardware?

No. The InnerWeb software update does not replace the client's readers or cards.

Accept updates with proof

We will provide an acceptance protocol template tailored to your plant. Book a conversation with InnerWeb.

Let's talk