Schedules and bandwidth
Each destination can be held to a pace that suits the link it sits behind: a remote site over a thin WAN, a cloud service over the hospital’s internet line.
A standing limit
Section titled “A standing limit”Bandwidth limit (Mbit/s) caps how fast each node sends to the destination. It applies per node: two nodes with a 10 Mbit/s limit can together send 20.
Schedule windows
Section titled “Schedule windows”Recurring windows, in each node’s local time, either pause sending or apply a lower limit:
| Days | From | To | Action |
|---|---|---|---|
| Mon–Fri | 07:00 | 19:00 | Limit 20 Mbit/s |
| (every day) | 02:00 | 04:00 | Pause (the destination’s backup window) |
Outside every window the standing limit applies (or none). Times are HH:mm; a window whose end is before its start
runs past midnight (22:00–06:00). The first window that matches applies.
- While paused, instances wait in the queue and are sent when the window ends. High-priority instances can still be sent during a pause (on by default), so urgent studies get through.
- A paused destination does not raise the oldest instance too old alert.
- Limits and windows apply at once, without reconnecting.
Typical uses
Section titled “Typical uses”- Send to an outside reading group’s site only at night, or slowly during the day.
- Keep a cloud archive’s uploads from crowding out the clinic’s internet line in office hours.
- Move a backlog to a new archive over a weekend.