Operations
Day-to-day operation of a SkyNode-paired telescope falls into a few modes:
- Automated observing — automation started, control authority
remote, no active lease. SkyNode pulls tasks from the hub and executes them. - Manual takeover via lease — short, remote operator intervention through the hub during otherwise automated nights.
- MANUAL control at the telescope — for maintenance, engineering, or manual observing nights.
- Safe shutdown — invoked deliberately or triggered by a shutdown interlock.
This page walks through each, with notes on what an owner is authorizing and what subsystems are involved.
Automated observing
Section titled “Automated observing”When the operator has started automation (Settings → Startup & Shutdown → Start automation; the choice survives restarts), the control authority is remote, the hub connection is up, and no manual control lease or shutdown interlock is active, SkyNode runs in its “normal” state. While any of these is missing, the top status bar shows automation as blocked with the reason. A simplified view of the processes involved:
| Subsystem | Role |
|---|---|
| Telescope controller | Top-level orchestration: open/close the session, spin up the task manager, handle lifecycle events |
| Task manager | Polls the hub for eligible tasks, claims them, and runs each through the appropriate task executor |
| Task executors | Per-observation-type — optical_imaging, optical_imaging_calibration, radio_mapping, radio_tracking. Drive the per-device drivers through one observation type’s logic |
| Target acquisition manager | Converts hub targets into ephemeris providers; keeps mount and enclosure aligned |
| Enclosure controller | Manages enclosure state (open/close, rotate to align with target) |
| Operational constraint evaluator | Continuously evaluates constraints; blocks task start when any required constraint fails |
| Resource manager | Device health, thermal, battery, storage — feeds into constraints and shutdown logic |
| State sources + router + stores | Aggregate snapshots from all subsystems into unified streams |
| Hub publisher | Publishes snapshots and events back to the hub |
| MQTT publisher | (Optional) Mirrors snapshots to the local broker |
For a given task, the executor walks a deterministic sequence of phases. The exact phase set differs by observation type but covers acquisition, configuration, capture, and post-processing.
SkyNode also uploads the frames it produces to the hub; the Files page in the local GUI shows each frame from acquisition through transfer.
Manual control leases
Section titled “Manual control leases”A manual control lease is an override taken through the hub. While it’s active, automation stops; once it expires or is released, automation resumes on its own.
Use a lease when you need a short remote takeover but want the telescope to return to automated observing without operator intervention.
Requesting a lease
Section titled “Requesting a lease”Take a lease from the telescope’s Control tab on the website, or through the API. Any user with update permission on the telescope can take it; there is one lease per telescope, and while someone else holds it a request is refused. The website takes a three-minute lease and renews it every minute while the page is open, so a closed page lets the lease lapse within minutes. Through the API the duration (ttl_seconds) is optional, and a lease taken without one does not expire until it is released or revoked.
To drive a telescope by hand at the site, use MANUAL control in SkyNode instead (see below).
Holding a lease
Section titled “Holding a lease”While the lease is active:
- The task manager stops: it picks up no new tasks, and a task that is running is aborted right away.
- The lease holder’s commands sent through the hub (from the website’s Control tab) are accepted, subject to the node’s own policies and interlocks.
The lease emits EventManualControlLease events on the snapshot stream so anyone watching (hub-side dashboards, local MQTT subscribers) sees the takeover in real time.
Releasing a lease
Section titled “Releasing a lease”The lease expires at its expires_at if it has one. The holder can release it from the website at any time. Automated task execution resumes within seconds of release.
Revoking a lease from the telescope
Section titled “Revoking a lease from the telescope”On the website only the holder can release a lease. The telescope’s own SkyNode can revoke any lease, whoever holds it: use Revoke on the lease in the SkyNode Overview — useful when the remote holder forgot to release it or can’t be reached.
MANUAL control at the telescope
Section titled “MANUAL control at the telescope”For operation at the site — engineering nights, maintenance windows, extended weather closures — switch SkyNode to MANUAL instead of relying on leases. The switch is the control authority in SkyNode’s top status bar (or the button the Controls page offers); it is always confirmed, and the choice survives restarts.
While SkyNode is in MANUAL:
- The task manager does not pull tasks at all.
- Remote commands are refused, even from a lease holder: the person at the telescope comes first.
- The hub still receives snapshots, log entries, and state, but no scheduled work runs.
- The Controls page offers the local operator the telescope’s devices: imaging, mount, focuser, filter wheel, rotator, enclosure open/close requests and mirror covers, and IP cameras.
Switch back to REMOTE when you want automated observing to resume. A command started locally does not survive the switch to REMOTE, nor a hub command the switch to MANUAL.
Manual commands
Section titled “Manual commands”Every manual action is a named command in one registry (mount.slew, mount.park, focuser.move, enclosure.request_open, …), run by one runner that checks whether it is allowed right now and supervises it. The local Controls page and the hub’s command channel both go through that runner, so a command from the website gets exactly the checks a local one does. There is no generic surface for arbitrary device calls.
Shutdown actions
Section titled “Shutdown actions”The configured shutdown actions (Configuration → controller_lifecycle) run in two situations: when a shutdown interlock fires (the lifecycle supervisor applies them until the interlock clears), and when the operator quits SkyNode and chooses to shut the observatory down first.
The quit-time sequence, in order (defaults in parentheses):
- Stop mount tracking (always).
- Park the mount (
park_mount: true). - Park the enclosure (
park_enclosure: false— rotate a dome to its configured park azimuth; for domes whose shutter needs a contact position). - Close the enclosure (
close_enclosure: true), through the enclosure controller, which keeps it closed afterwards. - Close the mirror covers (
close_covers: true). - Stop the mirror fans (
stop_fans: true). - Warm the cameras to ambient (
warm_cameras: false— off by default; enable for installations where rapid cool/warm cycling is hard on the sensors).
When a shutdown interlock fires, the same flags apply except the enclosure park, and the enclosure and covers are requested closed before the mount parks so the motions overlap. In MANUAL control an interlock holds the enclosure closed but does not park the mount from under the operator; the rest of the shutdown runs when the node returns to REMOTE with the interlock still active.
Edit these in Settings → Startup & Shutdown. Each step runs only for a device this SkyNode manages and the driver can do.
Quitting SkyNode
Section titled “Quitting SkyNode”Closing the window and installing an update go through one dialog. It shows what the mount, enclosure and cameras are doing and offers two choices:
- Shut down the observatory, then quit — the sequence above, listing the steps this node can run and, for any it cannot, why (an enclosure that is configured but not connected, a roof SkyNode can only observe, a mount that cannot park). This is the default whenever something is worth shutting down: the shutter is open, the mount is tracking or unparked, a camera is cooling.
- Quit and leave the hardware as it is — SkyNode only disconnects; the dialog says what stays as it is (the enclosure stays open, the mount keeps tracking) and that automation stops while SkyNode is closed.
The sequence runs step by step with progress and can be aborted; a dome that is closing keeps closing. A step that fails is shown before you decide whether to quit anyway or stay open; staying restarts the automation the sequence stopped.
While a shutdown interlock is active, the top status bar shows automation as shutting down or shut down, with the reason. You can also set a shutdown lock yourself from the Lock button on the Overview, and cancel it there.
Bringing the telescope back up after a shutdown
Section titled “Bringing the telescope back up after a shutdown”When a shutdown interlock clears, the lifecycle supervisor runs the configured startup actions before automation resumes — provided automation is started, the control authority is remote, no lease is held, and the node is operational:
- Unpark the mount (
unpark_mount: trueby default). - Restore cameras to their cooling setpoints (
restore_cameras: falseby default).
After startup actions complete, the task manager begins polling again automatically — no operator intervention needed.
Startup actions are not run when SkyNode itself starts. After a restart (clean or not), a node whose automation was started resumes it, and devices whose auto-connect is armed reconnect on their own; there is no separate recovery flow.