Forum Discussion

InquisitiveMai's avatar
InquisitiveMai
Icon for Cirrostratus rankCirrostratus
Aug 20, 2024

SNMP OIDs for Memory, CPU and Disk Usage

I have referred to few KB articles and Devcentral https://support.f5.com/csp/article/K6420. I compared files  /var/run/bigip_error_maps.dat /etc/alert/alert.conf for High CPU OID.Looks like the OID...
application delivery
CeinarF's avatar
CeinarF
Icon for Employee rankEmployee

Hello,

it should depend on TMOS version your devices are running, but I'm quite sure you can use the following OID.

CPU

  • .1.3.6.1.4.1.3375.2.1.1.2.1.41.0 sysStatTmTotalCycles.0
  • .1.3.6.1.4.1.3375.2.1.1.2.1.42.0 sysStatTmIdleCycles.0
  • .1.3.6.1.4.1.3375.2.1.1.2.1.43.0 sysStatTmSleepCycles.0
  • To calculate CPU utilization % take two samples, calculate the delta, and calculate: ((<DeltaTmTotalCycles> - (<DeltaTmIdleCycles> + <DeltaTmSleepCycles>)) / <DeltaTmTotalCycles>) * 100

Memory

  • .1.3.6.1.4.1.3375.2.1.1.2.1.44.0 sysStatMemoryTotal.0
  • .1.3.6.1.4.1.3375.2.1.1.2.1.45.0 sysStatMemoryUsed.0
  • To calculate memory utilization %: (sysStatMemoryUsed/sysStatMemoryTotal) * 100

Disk

  • .1.3.6.1.4.1.3375.2.1.7.3.2.1.3.1.47 sysHostDiskTotalBlocks."/"
  • .1.3.6.1.4.1.3375.2.1.7.3.2.1.4.1.47 sysHostDiskFreeBlocks."/"
  • Depending on how many partition are built, you need to use the same OID paths in each partition, for example .1.3.6.1.4.1.3375.2.1.7.3.2.1.1.4.47.117.115.114 is sysHostDiskPartition."/usr" and .1.3.6.1.4.1.3375.2.1.7.3.2.1.4.4.47.117.115.114 sysHostDiskFreeBlocks."/usr"

I don't get what do you mean with "OID for SNMP not reachable".

SNMP is transported via UDP, so is a connectionless protocol with no guarantee of delivery, ordering, or duplicate protection.

BR

InquisitiveMai's avatar
InquisitiveMai
Icon for Cirrostratus rankCirrostratus
Aug 21, 2024

Thank you for your response. You meant

  • To calculate CPU utilization % take two samples, calculate the delta, and calculate: ((<sysStatTmTotalCycles> - (<sysStatTmIdleCycles> + <sysStatTmSleepCycles>)) / <sysStatTmTotalCycles>) * 100

 

"SNMP not reachable", I was thinking if there is anyway a alert can be generated from F5. Will there be any log or any custom alert?  I am thinking with the monitoring tool if F5 is not reachable through SNMP it should be able to trigger a alert.

  • CeinarF's avatar
    CeinarF
    Icon for Employee rankEmployee
    Aug 21, 2024

    Sorry I pasted a wrong formula.

    OID doesn't give you the percentage of CPU utilization, so to calculate CPU utilization percent take two samples, calculate the delta, and calculate: ((<DeltaTmTotalCycles> - (<DeltaTmIdleCycles> + <DeltaTmSleepCycles>)) / <DeltaTmTotalCycles>) * 100

    Monitoring tool normally monitors F5 IP, so if the IP is not reachable the F5 become "red". Just like F5 have monitor on Pools/members/nodes used by LTM.

    If something is going wrong inside F5, TMOS generates internal alarm/logs, so can F5 can generate SNMP Trap or send "error log" to remote syslog servers or send email using an SMTP server.

    So if you're asking "How F5 is able to send alarm if something is going wrong?", you can use one (or more) of the 3 options (SNMP Traps, remote syslog, email using external SMTP server).

    BR