Notifications
Your server can tell you when something needs you: a computer waiting for approval, a satellite gone offline, a disk nearly full, an update to install. Its notifications appear on your spandock.com dashboard, under Notifications, so you see them without opening the server.
Notifications are in every plan. They are off until an administrator connects the server.
What a server reports
| Notification | When |
|---|---|
| A computer is waiting for approval | A new computer paired and waits for you |
| A server or satellite went offline | A satellite has been unreachable for 3 minutes |
| A server's disk is nearly full | On the server or a satellite |
| An update is available | A new release is out; critical and security updates are marked as such |
| The license is about to lapse | The license is in its grace period or past due, or the trial ends within 3 days |
| Spend went over a budget | Someone went over their model-proxy budget this period |
| Errors are rising | Errors are well above the usual level for the last 15 minutes |
| A guardrail blocked prompts or tool calls | Counts only, at most once an hour |
| Just-in-time access is waiting for approval | Someone asked for access, on Enterprise |
Each one is sent once when it starts, and again only if it ends and comes back.
What a notification contains
Only names of servers and computers, and counts. Never telemetry, prompts, code, tool output, file paths or what people did. A budget notification says "someone" and the count; you see who on the server's dashboard.
Turn them on
- On the server's dashboard, open Settings → Notifications and choose Connect. The server must be activated with your spandock.com account.
- For each kind of notification, switch Show on spandock.com on or off.
- Choose Send a test to check that it arrives.
Only the hub reports. Its satellites' state reaches it anyway, so you get one notification per event, not one per server. Disconnect stops it at any time.
On spandock.com
Dashboard → Notifications lists what your servers reported, newest first, with Unread and Mark all read. It also shows which servers are connected and what each one sends. The settings themselves stay on each server: it decides what to send.
Uptime
Dashboard → Servers shows how much of the time each of your servers was up: the last 24 hours, 7, 30 and 90 days, and a bar for each of the last 30 days. Open a server to see its satellites' uptime too. Uptime is in every plan.
It comes from the hub's heartbeat, which the hub sends every 5 minutes while Settings → Uptime → Report uptime to the dashboard is on (on by default). That works without connecting notifications. A heartbeat missing for 15 minutes counts as an outage. While the hub is down, its satellites' state is unknown and isn't counted. Days are in UTC, and the history is kept for 90 days.
The heartbeat carries each satellite's installation ID and whether it is online. Never names, addresses or data.