Skip to main content

Offline work and synchronization

SpaceRivers caches account-specific field data so supported work can continue during a network interruption. Cached data can include the employee profile, company preferences, jobs, phases, cost codes, rate codes, expense types, units, schedules, and recent time entries.

Supported offline queue​

The app can save these pending actions locally for the active user:

  • clock-in;
  • clock-out after its related clock-in has a server identity;
  • Other Time; and
  • Expense.

Each saved account has its own working set. The synchronizer skips queued items that do not belong to the currently active user.

Synchronization sequence​

flowchart LR
    LOCAL[Saved on device] --> IN[Send clock-in]
    IN --> ID[Receive server ID]
    ID --> OUT[Send related clock-out]
    LOCAL --> OTHER[Send Other Time or Expense]
    OUT --> REMOVE[Remove confirmed queue item]
    OTHER --> REMOVE

Pending actions are retried when Clock resumes and around supported clock operations. Return to Clock after reconnecting and keep the app open long enough for the request to finish. Refresh Home, Schedule, and Timesheet separately to load their latest server views.

Safe offline practice​

  • Confirm the app says an item was saved locally before leaving the screen.
  • Do not uninstall the app, clear storage, reset the device, or delete app data while work is queued.
  • Do not create an online duplicate while a local copy is pending.
  • Keep the same user active until that user's queue is synchronized.
  • After reconnecting, verify the record in Timesheet and confirm its expected status.

Partial clock synchronization​

An offline clock-in must first receive a server ID before its clock-out can be sent. If a switch or clock-out cannot finish, preserve the local data, reconnect, and retry. Ask an administrator to compare the mobile time with the web Time Editor before manually recreating anything.

Offline mode does not bypass validation, permissions, approval, or payroll processing. A queued record can still be rejected when the server evaluates it.