Typical scenario on the shop floor
The watch worn by the employee sends position and status signals to the system, and the watch application logs in to the server with a token. The access token has a limited lifetime, and the refresh token has a longer lifetime and can be used only once. For a regular user, an expired token means a login screen where a person types a password or code.
The watch has no keyboard. When the refresh failed, for example after a longer network break, the device had no way to log in again and went silent. On the dispatcher's side it looked like an ordinary loss of signal: the employee stopped refreshing on the map, and the cause was neither the network nor the battery, but the authentication state.
Why a kiosk device needs a separate rule
The system distinguishes a regular user account from the account of a kiosk device. A device without a keyboard cannot log in again after its token expires, so it needs a separate flag on the account that is switched on explicitly. Without it, standard rules apply: the token expires and access ends.
This design is deliberate. Removing the token lifetime limit for all accounts would weaken the security of the whole installation, while the flag covers only those accounts that IT deliberately marks as devices without a person at the keyboard.
Why migration does not set the flag on any accounts
The change that introduces the flag does not mark any account. Account identifiers differ between installations, and granting permissions during an update would be a silent change that the administrator might not notice. After the update, the plant's IT department indicates which accounts are kiosk devices and records this in its change register.
How to count the fleet and find devices that dropped out
Every admission and every refusal after the device token's grace period is written to the log with a warning level, account and expiry date. This gives two practical uses: the administrator sees which device accounts rely on the grace period, and can count how many devices in the fleet actually work under this rule.
A simple review after deployment means comparing the list of accounts with the flag against the warning entries from the last week. A flagged account with no entries and no signal needs a visit to the device. An unflagged account that appears in refusals is a candidate for flagging, if it really is a device without a keyboard.
- Collect the list of device accounts without a keyboard: watches, kiosks, permanently mounted tablets.
- Switch the flag on only for these accounts and record the decision in the IT change register.
- Every week, compare warning entries with the list of flagged accounts.
- React to an account that disappears from both the log and the map at the same time.
Token lifetimes in planning
The access token works for a shorter time and the refresh token for longer. When refresh happens regularly, the device has no problem. Trouble starts when the device is out of range or switched off for longer than the refresh window, and then, without the kiosk account flag, there is no way back.
For planning, assume that every break longer than the refresh window needs human intervention, and every shorter one should end with the device returning on its own once the network is restored. A watch charged once per shift usually fits within this window with a large margin.
What IT sets so the device returns on its own
The list of actions is short, but it needs agreement with health and safety and security, because the employee wears the device and data about their position matter for work safety. The IT department is responsible for the account, its flag, network policy and monitoring of the log.
If the installation uses a second login factor, kiosk devices cannot be blocked by a code requirement they cannot enter. Details of the settings are described in our recommendations for IT on MFA in the health and safety system.
- A separate account for each device, never a shared account for the whole fleet.
- Kiosk account flag only where there is no keyboard.
- Disable the MFA code requirement for device accounts, compensated by restricting addresses and permissions.
- Alert on missing signal from a device on the dispatcher's side.
Conclusions for the plant
A failure of a device without a keyboard is not a hardware defect, but a consequence of login security designed for a human. The solution is an explicit IT decision for each account, not a global relaxation of the rules.
A well-run deployment ends with a list of device accounts, a register of decisions and a simple weekly report from the log. That is a few minutes of work a week, which replaces night-time searches for the cause of a silent watch.
Video
Frequently asked questions
Does the kiosk account flag lower the security of the whole system?
No, because it covers only accounts indicated by IT. All other accounts keep standard token expiry and the requirement to log in again.
Will device accounts be flagged automatically after a system update?
No. Migration does not set the flag on any account, because identifiers differ between installations. The plant administrator indicates the accounts.
How do I check which devices use the token grace period?
The log records admissions and refusals with a warning level, account and expiry date. Comparing these entries with the fleet account list shows the usage.
Let's check how your devices log in
We will go through device accounts without a keyboard with your IT department and set a rule that lets them return to work on their own.
Let's talk