Alert description
This alert indicates that the oblogservice process does not exist (is not running) in the LogService cluster.
Alert principle
The following table lists the key parameters involved in the monitoring logic of this alert.
Parameter |
Value |
|---|---|
| Monitoring Metrics | logservice_process_exists: Indicates the status of the OBLOBS service process. 1 indicates the process is alive, and 0 indicates the process does not exist. |
| Monitoring Expression | process_exists{name="oblogservice",@LABELS} |
| Metric Collection | process_exists{name="oblogservice"} |
| Metric Source | OCP MonAgent collects process liveness status through the processInput plugin. |
| Collection Cycle | 1 second |
Rule information
Monitoring Metrics |
Default Threshold |
Duration |
Detection Cycle |
Elimination Cycle |
|---|---|---|---|---|
| logservice_process_exists | 0 (process does not exist) | 30 Seconds | 10 Seconds | 5 Minutes |
Alert expression: logservice_process_exists == 0
Alert information
Alert Trigger Method |
Alert Level |
Scope |
|---|---|---|
| Based on monitoring metric expression | Emergency | Host (LogService) |
Alert template
Alert overview
- Template: ${alarm_target} ${alarm_name}
- Example: logservice_cluster=my_ls_cluster:host=11.124.9.47 The LogService process does not exist.
Alert Details
- Template: LogService cluster: ${logservice_cluster}, host: ${host}, alert: ${alarm_name}.
- Example: LogService cluster: my_ls_cluster, host: 11.124.9.47, alert: LogService process does not exist.
Alert recovery
- After the process resumes running and its metric value returns to 1, it will automatically expire after the clearance period (5 minutes) expires.
Impact on the system
When the oblogservice process does not exist, the node is completely unavailable. The number of available replicas in the LogService cluster decreases, which may cause log service interruptions or degraded read/write performance.
Possible causes
- Process startup failed
- The process was manually stopped or crashed abnormally and did not automatically restart.
- Host failure or network unreachability
Solution
- Check whether the host is reachable and confirm that the installation directory of oblogservice is normal and permissions are correct.
- Manually start the oblogservice process and observe the startup logs.
- Perform node O&M operations (restart/replacement) through the OCP log service cluster management interface
- If the instance still cannot be started, contact technical support and provide the process logs.