Switch monitoring

[Valentin Richomme]

Hello Community

I have a problem monitoring switch stacks when I can monitor switches alone. Has anyone encountered this problem before or has a solution ?

Thank you in advance for your feedback

Valentin

[Tom Black]

What’s the difference for switch stack and switches?

Regards 

--


You received this message because you are subscribed to the Google Groups "Prometheus Users" group.


To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.


To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/3a1aecb8-c0bc-4047-905c-287620ac2396n%40googlegroups.com.

[Valentin Richomme]

a switch stack is several switches connected to each other forming a single switch at the logical level.

so normally there's no difference with a switch alone but I can't monitor them
the errors are either :

- server returned HTTP status 500 Internal Server Error

or either :

- Get "http://localhost:9116/snmp?module=if_mib&target=my_stack": context deadline exceeded

Regards

[Brian Candler]

On Wednesday, 9 September 2020 15:29:03 UTC+1, Valentin Richomme wrote:

- Get "http://localhost:9116/snmp?module=if_mib&target=my_stack": context deadline exceeded

You haven't mentioned the vendor or model.  It sounds like a bug with your device, or it's taking too long to respond because it has so many interfaces.  (I had this problem with old Dell Powerconnect 5524 switches; they show hundreds of interfaces as if they were in a stack.  I could scrape them at 1 minute intervals, but not at 15 second intervals.  The newer Powerconnect N1548 are much faster)

What happens if you use snmpbulkwalk to walk the IF MIB of the stack?  How long does it take to complete?

You could also try

tcpdump -i eth0 -nn udp port 161 and host my_stack

while you're doing a scrape, to see if data is trickling back as it runs.