Forum Discussion
smp_86112
Cirrostratus
CirrostratusHow do you monitor your LTMs/GTMs
We wrote some perl scripts that leverage iControl which we execute with Nagios to check the status of our F5 units. But the frequency with which we execute those checks may be causing us problems.
I'd like to get some input on how you make sure your LTMs/GTMs are functioning OK. What tool do you use, and how frequent do you check?
hoolioI expect you'll get many different suggestions/opinions on this. Here are some quick thoughts:
Many customers I work with use external monitoring tools or services that make requests to the LTM VIPs or GTM WIPs. Ideally they're checking dynamically generated content that touches the database to track client experiences. This is one important view of the entire health of the application stack including the F5's.
You can also monitor CPU/memory usage via SNMP or CLI tools. For spot checks from the CLI standard Linux tools help, along with tmstat.
Aaron
smp_86112I've spoken to several sources who all say SNMP should be the standard. Taking that a bit further, what OIDs are people querying to determine the overall health of the system? I'm not interested in the OIDs relating to TMM - I have other ways of validating that our application traffic is being processed. I'm more concerned about the system health.- The_Bhattman
Nimbostratus
I suppose what you would want to check is the following
CPU
MEMORY
DISKSPACE
Processess
bigd
mcp
named etc
These are the elements to check for the overall health of the BIGIP.
You will not find SNMP OID for process - at least I couldn't find it so hopefully your monitor can do more then simply SNMP checking.
I hope that helps
Bhattman
