Monitoring CPU load
| SNMP object: | HOST-RESOURCES-MIB::hrProcessorLoad |
CPU load shows the availability of a CPU. Its value is a rational number ranging from 0.0 or 0% to 1.0 or 100%. If the CPU load is 0.0 or 0%, then the measured CPU is idle. If the value is 1.0 or 100%, then the CPU is fully loaded.
If the load of one CPU is above 90% and other threads are significantly less loaded for a longer time period, fine-tune your configuration or purchase more SSB appliances. For assistance, contact our Support Team.
Monitoring CPU load averages
| SNMP object: | UCD-SNMP-MIB::laLoad |
CPU load averages (or system load averages) is the average load of SSB's CPUs and the size of the task queue, during the past 1, 5, and 15 minutes, respectively.
If the load is constantly equal to or higher than the number of CPUs in your appliance, fine-tune your configuration or purchase more SSB appliances. For assistance, contact our Support Team.
Monitoring CPU usage
User CPU time:
| SNMP object: | UCD-SNMP-MIB::ssCpuUser |
If processor is not in idle (this means that there is live log traffic), it is quite normal that the majority of the CPU time is spent on running user space processes.
System CPU time:
| SNMP object: | UCD-SNMP-MIB::ssCpuSystem |
The amount of time spent in the kernel should be as low as possible. Ideally, around 0.5% of the time given to the different processes is spent in the kernel. This number can peak much higher, especially when there are a lot of I/O activities.
Idle CPU time:
| SNMP object: | UCD-SNMP-MIB::ssCpuIdle |
The total of the user CPU time percentage and the idle CPU percentage should be close to 100%. If the CPU spends a lot more time in other states, it is worth investigating the root cause, because it can indicate issues.
Monitoring the details of your CPU usage:
To fine-tune monitoring your CPU usage, you can use the following values. It is also possible to generate a chart from these values.
| Raw CPU user time | UCD-SNMP-MIB::ssCpuRawUser |
| Raw CPU nice time | UCD-SNMP-MIB::ssCpuRawNice |
| Raw CPU system time | UCD-SNMP-MIB::ssCpuRawSystem |
| Raw CPU idle time | UCD-SNMP-MIB::ssCpuRawIdle |
| Raw CPU wait time | UCD-SNMP-MIB::ssCpuRawWait |
| Raw CPU kernel time | UCD-SNMP-MIB::ssCpuRawKernel |
| Raw CPU interrupt time | UCD-SNMP-MIB::ssCpuRawInterrupt |
| Raw CPU Soft IRQ time | UCD-SNMP-MIB::ssCpuRawSoftIRQ |
| Raw CPU steal time | UCD-SNMP-MIB::ssCpuRawSteal |
| Raw CPU guest time | UCD-SNMP-MIB::ssCpuRawGuest |
| Raw CPU guest nice time | UCD-SNMP-MIB::ssCpuRawGuestNice |
Monitoring SSB's IO
Disk IO per partition
| SNMP object: | UCD-DISKIO-MIB::diskIOTable |
-
sda
If the 15-minute load (for details, see Monitoring CPU load averages) is getting close to 90%, your system does not have enough resources and you probably need to purchase more SSB appliances. For assistance, contact our Support Team.
-
sdb
NOTE: This is only available on SSB T1 appliances.
If the 15-minute load (for details, see Monitoring CPU load averages) is getting close to 90%, your system does not have enough resources and you probably need to purchase more SSB appliances. For assistance, contact our Support Team.
Interfaces IO by interface name
| SNMP object: | RFC1213-MIB::ifTable |
The following interfaces can be monitored (the type of traffic that can affect the load):
-
eth0 - external (network, redundant HA, next hop monitoring)
-
eth1 - management (network, redundant HA, next hop monitoring)
-
eth2 - internal (redundant, next hop monitoring)
-
eth3 - HA
If the load on an interface seems to be too high, check whether you have configured SSB in a way that affects that node. For example, if you do not use a management interface, the load on the external interface can be higher. Or, configuring next hop monitoring can also increase the load on an interface.