Monitoring

Long-running experiments need more than the main data stream. Quatm monitors the laboratory infrastructure and the health of its own services.

Environmental monitors

Bundled monitors cover:

• temperature controllers and sensors;
• magnetometers / Hall probes;
• wavemeters and laser power meters;
• vacuum gauges;
• PC health (CPU, memory, disk);
• oscilloscopes and general SCPI instruments.

Monitors are ordinary services that publish into the property and data streams. A failed monitor is reported and can be restarted without touching the experiment runtime.

Device health

The Device Manager can probe instruments, run driver self-tests and show the current lease owner. Health state and history are kept separately from experiment data, so a failing instrument does not corrupt a run’s HDF5 file.

Service health

The Operations Console and Task Monitor show:

• service process state and uptime;
• task queue state and heartbeats;
• configured endpoints and reachability;
• data-manager save errors and disk headroom;
• resource alerts (disk, database size, threads, file descriptors).

Alerts are rate-limited and transition-based: an ongoing error is reported once, and recovery is reported separately.