InnerWeb
Skip to main content

IT department FAQ before a safety system rollout: network, accounts, email and logs

Before a safety system goes live, the IT department asks the same questions: where it runs, what goes out to the internet, who creates accounts and what stays in the logs. We collected answers based on the feature documentation, marking what is standard and what is an option to agree with the plant.

7 days
trusted device memory
Source: card
107
entities covered by the audit trail
Source: card
10 min
validity of the MFA code from email
Source: card

Where does the system run and who maintains it?

Each customer has an isolated instance with its own database and its own email system. The instance runs in Docker containers, also on the customer's server, and some environments run on Kubernetes. The data of one plant does not mix with the data of another, so the IT department assesses the risk for one clearly limited installation.

The configuration is described by an environment variable contract without values: a list of names the instance accepts and a description of their meaning. Values, including secrets, are set in the runtime environment and not in the code repository. For IT this means one simple point to agree: who stores the values, where, and how they are rotated.

Does the application host need internet access?

Not always. When the plant network blocks all outbound traffic from the application host, alarm webhooks are stored durably in the system, and an external poller picks them up over an inbound channel. An alarm is not lost just because the firewall did not allow an outbound connection.

For IT this leads to a network architecture choice. The host can be left without outbound traffic and receive alarms on the inbound side, or a narrow rule can be opened for specific addresses. Both paths are acceptable, but it is worth recording the decision before go-live, because it determines the scope of alarm tests.

Scope of the audit trail after the review

Scope of the audit trail after the reviewBefore and after chart: the audit trail scope grew from 60 to 107 entities0255075100125entitiesBefore the review: Entities in the audit: 60 entities60Before the reviewAfter the review: Entities in the audit: 107 entities107After the review
Source: Cell card: Audit Trail feature documentation

How does login work and can MFA be enabled?

Multi-factor authentication offers two factors: a 6-digit code sent by email, valid for 10 minutes with a limit of 5 attempts, and TOTP compatible with Google Authenticator. MFA is an option enabled per installation in the security settings and is off by default, so the plant makes the decision whether to turn it on.

A trusted device remembers the verification for 7 days thanks to a cookie. This affects how many codes users request in a week, and therefore also the helpdesk workload. We gathered the recommendations for IT on a separate MFA page, and the cost of login support is included in the sample calculation.

Does the system feed accounts from the plant's user directory?

The feature documentation does not describe feeding accounts from the plant's user directory. The described SSO covers entry from the qualification portal with a person identifier in the id_token. Directory integration is therefore a topic to agree with the IT department, not a standard that can be assumed in the schedule.

In practice: IT prepares the account list and indicates the source of truth about people, and the rollout team indicates how the accounts are loaded. We describe this as an on-request option, with the scope agreed before go-live. A guide on preparing the account list is on a separate page.

How are the API and portal integrations secured?

The API for the qualification portal requires a machine token in the header. A missing or wrong token returns a 401 response. Search queries require at least 3 letters or digits and return at most 8 results, which limits the risk of mass data download.

Integration keys do not go to the browser. The SSO panel shows settings in read-only mode, and approving, rejecting and re-issuing a decision on an application is limited to one administrative role. When registration from a QR poster is enabled, the same token creates accounts and companies and issues passes, so the risk assessment and the token rotation plan should cover this scope.

What stays in the logs and how does the system report errors?

The audit trail covers entities changed by a person, and entities written by the machine are excluded, because their history is source data. In the review the scope grew from 60 to 107 entities. Timestamps use the Europe/Warsaw time zone from October 2026, and earlier entries were not recalculated.

Errors go to one shared Sentry project with a limit of 21 events per hour per error type and 89 per hour per instance for log entries. Security alarms are not silenced by this limit. The transport has a limit of 1 second per connection and 2 seconds per response, so an unavailable receiver does not prolong error handling.

What does outgoing email look like and do SPF records need changing?

The shared sending address of the system sits in the InnerWeb domain, and the SPF, DKIM and DMARC records remain unchanged. Only the displayed sender name and the Reply-To address change. The person appears under their own first name and company with a note via InnerWeb, and the system as System InnerWeb. The first name and company are shortened to 34 characters.

System messages, such as password reset, MFA code and password change confirmation, have no Reply-To. The plant's email may flag messages as suspicious when the sender comes from an external domain, so a good practice is to add the sending domain to the trusted list and inform users in advance. We describe the details in the material on sending email.

Video

Frequently asked questions

Does IT have to open outbound ports from the host?

It is not necessary. With blocked outbound traffic, alarm webhooks are stored durably and an external poller picks them up over an inbound channel. The architecture decision is recorded before go-live.

Is MFA enabled by default?

No. MFA is an option enabled per installation in the security settings. The plant decides whether to enable it and may leave trusted devices at 7 days.

Will the system load accounts from the user directory by itself?

The feature documentation does not describe this. Feeding accounts from the directory is agreed with the IT department as an on-request option, with the scope recorded before go-live.

Do SPF, DKIM and DMARC records need changing?

No. The shared sending address sits in the InnerWeb domain and the records stay unchanged. Only the displayed sender name and the Reply-To address change.

Sources

Let us go through the IT questions together

Let us arrange a short meeting with your IT department: we will go through the network, accounts, email and logs and record the decisions before go-live.

Let's talk