[j-nsp] QFX5100 System Process SNMP monitor

[Dale Shaw]

Hi Chris,

On Sun, 20 May 2018 at 1:01 pm, Chris Lee via juniper-nsp <
juniper-nsp at puck.nether.net> wrote:
>
> Hi all,
>
> We recently hit a jdhcpd bug in our QFX5100 VC (14.1X53-D30 release) which
> looks to be from the number of defunct zombie processes increasing over
> time leading up to an ungraceful failover of the routing engines.
>
> I have LibreNMS monitoring the QFX and it automagically graphs the running
> process count, but I'm struggling to figure out an SNMP MIB object number
> that gives me the same process count, as I'd like to monitor the same
value
> in our existing PRTG installation to send email/SMS alerts when certain
> thresholds are reached until we can follow JTAC's recommendation to
upgrade
> the QFX release to D46.
>
> I've downloaded the MIB pack from Juniper and tried grepping the files for
> "process" and "processes" but can't seem to find anything relevant.
>
> Anyone familiar with the SNMP on the QFX know where I might find total
> system process count ?

Junos implements HOST-RESOURCES-MIB, so you should be able to poll the
"hrSystemProcesses" object (gauge). Maybe that's how LibreNMS is doing it?

admin at gw> show snmp mib walk hrSystemProcesses
hrSystemProcesses.0 = 125

admin at gw> show system processes summary
last pid: 40910;  load averages:  0.20,  0.19,  0.17  up 29+08:45:43
08:45:48
126 processes: 18 running, 95 sleeping, 1 zombie, 12 waiting
[...]

Cheers
Dale


>
>
> Thanks,
> Chris
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp