harvest 24.11.1 netport.yaml ifgrp query removed

[wagneradrian92]

hey everyone,

i was just wondering why the query for the ifgrps got removed from the netport.yml config?
in version 24.5.0 this was still there:

• query: api/private/cli/network/port/ifgrp
  counters:
  • ^^ifgrp => port
  • ^^node => node
  • ^activeports => activeports
  • ^distr_func => ifgrp_func
  • ^down_ports => down_ports
  • ^ports => ports
  • ^up_ports => up_ports

is there a new way to collect metrics for ifgrps?

thanks!

regards
adrian

[Hardikl]

@wagneradrian92

Please refer this detail: #3117 (comment)

As mentioned in above link, it doesn't seems 1-1 mapping between call (api/network/ethernet/ports and api/private/cli/network/port - they are fine with same key and can add labels appropriately) and api/private/cli/network/port/ifgrp (totally different object call).

Hence, we have decided internally in team to remove this rest call as it's already showing the wrong data.
If required later, we can add new template for monitoring the ifgrp object separately.

Please open new feature request for monitoring the ifgrps object.

[wagneradrian92]

@Hardikl
Thanks for the explanation. i added the query to our config since we are alerting on the "activeports" attribute. waring if partial and alert when both ports of an ifgrp are down.

[Hardikl]

@wagneradrian92 ,

Could you please share the response of these below harvest rest cli commands, so we can replay them locally to understand better. Feel free to email them if you would prefer not to post them.

Note: Replace POLLER_NAME with your poller which has the required data of ifgrp.

./bin/harvest rest show data -p POLLER_NAME --api network/ethernet/ports > public_ethernet_ports.json

./bin/harvest rest show data -p POLLER_NAME --api private/cli/network/port --fields node,port,duplex_admin,duplex_oper,flowcontrol_admin,role,speed_admin" > private_ethernet_ports.json

./bin/harvest rest show data -p POLLER_NAME --api private/cli/network/port/ifgrp  --fields "ifgrp,node,activeports,distr_func,down_ports,ports,up_ports" > private_ifgrps.json

Also, If you could provide more information about how the activeports label or any other label would be used in your alerts, we would have a better understanding of your use case and provide best solution for it.

[wagneradrian92]

@Hardikl,
i sent you the files via email.
the alert we are currently using looks like this:

alert:NetAppIfgrpHealthDegraded
expr:netapp_net_port_status{activeports="partial",port_type="lag"}

so we simply check if an ifgrp has a degraded state and from there on check on the system where the problem may be. but since an ifgrp is redundant this is no "critical" alert for us so we have enough time to investigate the issue.

[Hardikl]

Thanks @wagneradrian92 for the json responses and alert rules about how it's consumed.

We have added the ifgrp api as part of the PR: #3342

[Hardikl]

Hi @wagneradrian92,
Changes are available in nightly.
Please give it a try when you get a chance and let us know if it solves the issue.