Public network probing, Agent heartbeats, and business services each have their own status. Once the anomaly is identified, you can initiate recovery from the node page, reusing the current firmware, identity, and configuration.
Understanding health layersThe platform probes the reachability of public network TCP/SSH and displays the probe results and timestamps. Whether the management entry point is reachable does not indicate whether business services are functioning normally.
View the most recent heartbeat and the reporting times for CPU, memory, storage, and resources. Node reporting and active public network probing are displayed separately.
Ingress and egress services have their own health results and error causes. An abnormality in a single service is not simply described as the entire node being unreachable.
With resource maintenance permissions, you can initiate "Diagnose and recover" for eligible nodes. The task checks the maintenance connection and firmware identity, recovers abnormal services, and restarts the Agent, preserving node identity, firmware version, existing configuration, and metering status.
The node must be fully installed, with SSH credentials and trusted host fingerprints registered; version locking, parallel upgrades, or an existing recovery task will restrict the operation. If the initial installation fails, enter diagnosis and recovery from the original installation task.
The associated Relay may be briefly interrupted. Leaving the node page will not interrupt an independent task; come back to check progress and specific results.
After recovery, confirm the task results, heartbeat and service health on the node page, then let the affected devices reconnect to the business.
If you need to change the firmware version, enter the upgrade process separately. The node details list the current running version and the target version, making maintenance easier to plan.
Learn about firmware upgrades →No. The management entry point, Agent heartbeat and data-plane services each have their own results. Even when the public network is reachable, ingress or egress services may still be abnormal, so you should check the corresponding service reports.
An abnormal initial installation should be diagnosed and recovered from the original installation task. Node service recovery is intended for nodes that have completed installation and meet the maintenance conditions; it does not bypass firmware identity, version locking or concurrent task limits.
No. The task runs independently; return to the node page to check progress and health review. Recovery uses the current firmware; if you need to change the version, submit an upgrade separately.
