HP-UX Job Queues (Load Average)
Job Queues
The job queue monitor checks the average number of processes that have been waiting for CPU and performing disk I/O over the last 1, 5, or 15 minutes. A value of 4 in /system/jobQueue5Min means that at the time of polling there was an average of 4 jobs in the queue over the last 5 minutes.
The MIB variables computerSystemAvgJobs1, computerSystemAvgJobs5, and computerSystemAvgJobs15 from the hp-unix MIB provides the resource value to the monitor.
To verify the load averages on the system, use the uptime (1) command.
Table 7-3 Interpreting Job Queues
Resource Name |
Value Range |
Interpretation |
---|---|---|
/system/jobQueue1Min |
integer |
average number of jobs in the queue in the last minute |
/system/jobQueue5Min |
integer |
average number of jobs in the queue in the last 5 minutes |
/system/jobQueue15Min |
integer |
average number of jobs in the queue in the last 15 minutes |
The minimum polling interval is 30 seconds. Unless your system load tends to fluctuate wildly and need load-balancing attention frequently, set a polling interval greater than or equal to the job queue interval: 1, 5, and 15 minutes, respectively.
Table 7-5 Examples of System Resource Requests
To be alerted when… |
Resources to monitor | Monitoring Parameters | |||
---|---|---|---|---|---|
Notify |
Value |
Option |
|||
fewer than 5 users are on the node, for running backups |
/system/numUsers |
when value is |
< |
5 |
|
more than 20 users are on the node, for load balancing, and when value returns to below 20 users |
/system/numUsers |
when value is |
> |
20 |
RETURN |
system load is high |
/system/ |
when value is |
> |
7 |
INITIAL |
/system/ |
when value is |
> |
4 |
INITIAL |
|
/system/ |
when value is |
> |
3 |
INITIAL |
|
file systems are running out of space |
/system/filesystem/availMb for: |
when value is |
< |
50 |
INITIAL |