network::f5::bigip::snmp::plugin --mode=tmm-usage: wrong and incomplete memory infos returned

[joschi99]

Latest plugin version

The actual plugin use the OID .1.3.6.1.4.1.3375.2.1.8.2.3 to monitor the TMM memory, but refering the official documentation from F5 https://my.f5.com/manage/s/article/K44014003 should used the OID .1.3.6.1.4.1.3375.2.1.1.2.21:

There are 2 memories and both should be monitored.

[lucie-dubrunfaut]

Hello :)

I'm not sure what you mean by there being 2 memories technically speaking? Within the same equipment?
It would be a matter of adapting the plugin so that it returns the same data as at present, plus the OIDs you mentioned:
.1.3.6.1.4.1.3375.2.1.1.2.20.44.0
.1.3.6.1.4.1.3375.2.1.1.2.20.45.0
.1.3.6.1.4.1.3375.2.1.1.2.21.28.0
.1.3.6.1.4.1.3375.2.1.1.2.21.29.0
Or return these OIDs if the ones currently being processed are not present?

[joschi99]

Hi @lucie-dubrunfaut,

unfortunatly i'm not a F5 specialist, but refering the F5 documentation there are 3 memory pools that should be considered to monitor:

1. TMM memory
2. Other memory
3. Swap memory

Here you can find some real value:

.1.3.6.1.4.1.3375.2.1.1.2.21.28.0 = Counter64: 15137243136
.1.3.6.1.4.1.3375.2.1.1.2.21.29.0 = Counter64: 516077576
.1.3.6.1.4.1.3375.2.1.1.2.20.44.0 = Counter64: 6894288896
.1.3.6.1.4.1.3375.2.1.1.2.20.45.0 = Counter64: 4731965440
.1.3.6.1.4.1.3375.2.1.1.2.20.46.0 = Counter64: 1073737728
.1.3.6.1.4.1.3375.2.1.1.2.20.47.0 = Counter64: 7213056

So regarding this documentation the returned values from the plugin are on other OID's and some memory is completly missing.

Please let me know if you need something.

[lucie-dubrunfaut]

Hello :)

Thank you for data test 🙏
After some research, it seems that there is a redundancy (difficult to determine whether it was intended by the manufacturer or not) in the TMM memory OIDs. According to the descriptions of the OIDs, there is an exact similarity between :
1.3.6.1.4.1.3375.2.1.8.2.3.1.31 and 1.3.6.1.4.1.3375.2.1.1.2.21.28

The total memory available in bytes for TMM (Traffic Management Module).
Use sysTmmStatMemoryTotalKb for gauge type.

https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.8.2.3.1.31
https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.1.2.21.28
as well as
1.3.6.1.4.1.3375.2.1.8.2.3.1.32and 1.3.6.1.4.1.3375.2.1.1.2.21.29

The memory in use in bytes for TMM (Traffic Management Module).
Use sysTmmStatMemoryUsedKb. for gauge type

https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.8.2.3.1.32
https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.1.2.21.29

---

As the other two memory OIDs are not dedicated to TMM, it would be relevant to monitor them, but perhaps in a dedicated memory.pm mode?
https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.1.2.20.44
"The total other non-TMM memory in bytes for the system.
Use sysGlobalHostOtherMemTotalKb for gauge type."
https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.1.2.20.45
"The other non-TMM memory in bytes currently in use for the system.
Use sysGlobalHostOtherMemUsedKb for gauge type."
https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.1.2.20.46
"The total swap in bytes for the system.
Use sysGlobalHostSwapTotalKb for gauge type."
https://oid-rep.orange-labs.fr/get/1.3.6.1.4.1.3375.2.1.1.2.20.47
"The swap in bytes currently in use for the system.
Use sysGlobalHostSwapUsedKb for gauge type."

---

Both should be fairly easy improvements. Knowing that in the first case on redundant OIDs, I would tend to think that the OIDs of the current plugin are certainly compatible with other versions of equipment than yours which means that your model is not currently managed at this level.