Reports and division of responsibility
IT teams most often ask where the software vendor's responsibility ends and the plant's begins. InnerWeb is the producer of software that integrates with the client's existing systems: access control, time and attendance, HR and ERP. This gives a simple split: the software, its updates and its integrations are the vendor's responsibility, while readers and access control cards remain the property and responsibility of the plant's existing supplier.
The system is offered on a subscription with a one-time implementation fee, and a pilot runs for 3 to 6 months before full rollout. During this period the plant's team learns the reporting process, which then works without changes in day-to-day maintenance.
What should a report contain?
A report should state the system version, the moment the problem occurred, the steps that lead to the error and a description of the effect the user saw. The version is key, because in error reports it is the short hash of the deployed image commit, the same as the image tag in the registry, so every error can be linked to a specific change.
Does a cache failure silence safety alarms?
No. Work safety alarms are not silenced by a cache failure: if a failure occurs, the alarm still rings and the log records the problem. This rule is deliberate, because a silent system in a hazard situation would be more dangerous than a false alarm.
Updates: how they are checked before release
Every change in the repository goes through static analysis, style checks and unit tests, and merging is blocked until the result is green. This is the first gate, common to all changes, including minor fixes.
The second layer is browser end-to-end (E2E) tests: the package covers 242 tests in 31 files and runs a real Chromium browser against a real backend. The tests replicate a user: login, the admin panel, work planning, work permits, reports, documentation, maps and analytics. This way they catch errors visible to users that unit tests miss, because unit tests run against a mocked API.
Does an update require replacing readers or cards?
No. An update on the InnerWeb side does not replace the client's readers or cards, because the system receives events from the existing access control. A software version change therefore does not touch the plant's physical infrastructure.
How long does the full test suite take?
The full E2E suite takes about 2 to 2.5 hours. It is run manually on request and automatically at midnight on weekends, without blocking merges. Parallel mode shortens the time almost linearly, but each test instance needs its own database and its own Redis database.
Duration of a full E2E test run
Error monitoring: what the vendor sees and what the plant sees
All deployed instances report uncaught exceptions to a shared error monitoring service, so the vendor usually sees a problem before the plant reports it. Reports are grouped by error type, and each one is linked to the release version, so the support team immediately knows which change it concerns.
Exceptions describing invalid input, such as no access, a missing page or a malformed request, are excluded from reporting as non-failures. An operator who types a wrong number or clicks a non-existent address therefore does not generate a failure report, and the IT team is not flooded with noise.
The overriding principle concerns safety: work safety alarms are not silenced, and any problem with the monitoring itself is recorded in the log, not in the behaviour of the alarm. The plant may ask to review the list of reports during the maintenance review, and the SLA parameters (availability, response and repair times) are set in the maintenance contract, for example availability of 99.5% per month.
Development and custom changes on order
The client asks whether, after go-live, it can order its own feature. The documentation lists the maintenance options: server and licences under an SLA, maintenance of transmitters with battery replacement as part of the service, and rental of mobile devices, billed as a monthly subscription. Custom feature development is agreed separately, with its scope and deployment window.
Every ordered feature follows the same control and testing path as general changes. The details of this path are described in the FAQ on custom feature development and in the recommendations on regression tests before an update.
Video
Frequently asked questions
Does a software update change anything in the card readers?
No. InnerWeb does not manufacture or replace access control readers. The system receives events from the existing access control, so the update concerns only the software.
How does the vendor know which change caused an error?
The release version in an error report is the short hash of the deployed image commit, the same as the image tag in the registry. Each error is therefore linked to a specific change.
Do user errors end up as failure reports?
No. Exceptions describing invalid input, such as no access or a malformed request, are excluded from reporting as non-failures.
How long is the pilot before full rollout?
From 3 to 6 months. After it, the system moves to full rollout under a subscription model with a one-time implementation fee.
Let's talk about system maintenance
We will explain how reporting, updates and development of the system work in your plant. Book a conversation with the InnerWeb team.
Let's talk