Architecture
Telemetry and Monitoring at Computle
Computle monitors its infrastructure and the workstations it hosts around the clock. Monitoring covers performance, availability and hardware health, so that faults are found and worked on before users report them.
Real-time monitoring
Section titled “Real-time monitoring”The agent on each workstation and the hypervisor under it report:
- Memory: total, free and used.
- CPU: count and utilisation.
- Disk: partitions, usage and file-system health.
- GPU: utilisation, temperature, power draw and memory use.
- Network: interface state, traffic, latency and packet loss.
- Storage health: continuous checks on the hypervisor’s drives.
Telemetry data stays within your tenant and is used only to operate the service.
Anomaly detection
Section titled “Anomaly detection”Alerts are raised on:
- unusual spikes in CPU, memory or disk use;
- sudden drops in network throughput;
- hardware faults such as disk read or write errors or a failing network interface.
Alerts go to the engineering team to assess. Historical trends are reviewed to catch gradual degradation and plan maintenance.
Network monitoring
Section titled “Network monitoring”Switches, routers and other network equipment at every site are monitored for uptime, bandwidth and hardware status. A failing component is replaced before it causes an outage where possible.
Outage detection and response
Section titled “Outage detection and response”If a workstation stops responding, monitoring detects it and recovery starts automatically: a running workstation that has gone quiet is restarted. Engineers and, where it affects them, customers are notified of significant performance or hardware issues. The current service status is published at status.computle.com.