Need more network port details to give Ethernet parity with FC

[jgasher]

Is your feature request related to a problem? Please describe.
When building dashboard for Ethernet interfaces there aren't as many details available as FC ports. Unfortanately the nic_common object doesn't have the same counters available as the fcp_port object.

Describe the solution you'd like
User the net-port-get-iter API to gather the following metrics:
[administrative-duplex]
[administrative-flowcontrol]
[administrative-speed]
[broadcast-domain]
[ifgrp-distribution-function]
[ifgrp-mode]
[ifgrp-node]
[ifgrp-port]
[ignore-health-status]
[ipspace]
[is-administrative-auto-negotiate]
[is-administrative-up]
[is-operational-auto-negotiate]
[link-status]
[mtu]
[mtu-admin]
[node]
[operational-duplex]
[operational-flowcontrol]
[operational-speed]
[port]
[vlan-id]
[vlan-node]
[vlan-port]

[vgratian]

Hi @jgasher, thanks for bringing this up. Feel free to create and add a template for the zapi collector on your own. Check the documentation for some guidance and examples.

You just have to translate your list of metrics into a subtemplate that Harvest understands. E.g. I think something like this should work:

name:            NicPort
query:            net-port-get-iter
object:           nic_port

counters:
  net-port-info:
    - ^administrative-duplex
    - ^administrative-flowcontrol
    - administrative-speed
    - ^broadcast-domain
    - ^ifgrp-distribution-function
    - ^ifgrp-mode
    - ^ifgrp-node
    - ^ifgrp-port
    - ^ignore-health-status
    - ^ipspace
    - ^is-administrative-auto-negotiate
    - ^is-administrative-up
    - ^is-operational-auto-negotiate
    - ^link-status
    - mtu
    - mtu-admin
    - ^^node
    - ^operational-duplex
    - ^operational-flowcontrol
    - operational-speed
    - ^^port
    - ^vlan-id
    - ^vlan-node
    - ^vlan-port

[vgratian]

I tested with a more refined subtemplate. Created conf/zapi/cdot/9.8.0/net_port.yaml:

name:                     NetPort
query:                     net-port-get-iter
object:                    net_port

counters:
  net-port-info:
    - ^administrative-duplex                   => admin_duplex
    - ^administrative-flowcontrol              => admin_flowcontrol
    - ^administrative-speed                    => admin_speed
    - ^broadcast-domain                        => broadcast
    - ^ifgrp-distribution-function             => ifgrp_func
    - ^ifgrp-mode                              => ifgrp_mode
    - ^ifgrp-node                              => ifgrp_node
    - ^ifgrp-port                              => ifgrp_port
    - ^ignore-health-status                    => ignore_health
    - ^ipspace                                 => ipspace
    - ^is-administrative-auto-negotiate        => admin_auto_negotiate
    - ^is-administrative-up                    => admin_up
    - ^is-operational-auto-negotiate           => op_auto_negotiate
    - ^link-status                             => status
    - mtu                                      => port_mtu
    - mtu-admin                                => port_admin_mtu
    - ^^node                                   => node
    - ^operational-duplex                      => op_duplex
    - ^operational-flowcontrol                 => op_flowcontrol
    - ^operational-speed                       => op_speed
    - ^^port                                   => port
    - ^vlan-id                                 => vlan_id
    - ^vlan-node                               => vlan_node
    - ^vlan-port                               => vlan_port

plugins:
  LabelAgent:
    value_mapping: status status up `0`

export_options:
  include_all_labels: true

Started poller, here is example output:

net_port_status{datacenter="MUCCBC",cluster="jamaica",broadcast="FabricPool",status="up",node="jamaica-01",ignore_health="false",ipspace="FabricPool",admin_up="true",port="a0b-59",vlan_id="59",vlan_node="jamaica-01",vlan_port="a0b"} 0
net_port_port_mtu{datacenter="MUCCBC",cluster="jamaica",ignore_health="false",ipspace="Default",admin_flowcontrol="full",op_auto_negotiate="true",op_speed="1000",broadcast="Cluster-MGMT_v59",admin_auto_negotiate="true",admin_up="true",admin_speed="auto",status="up",node="jamaica-01",op_duplex="full",op_flowcontrol="full",port="e0M",admin_duplex="auto"} 1500
net_port_port_admin_mtu{datacenter="MUCCBC",cluster="jamaica",ignore_health="false",ipspace="Default",admin_flowcontrol="full",op_auto_negotiate="true",op_speed="1000",broadcast="Cluster-MGMT_v59",admin_auto_negotiate="true",admin_up="true",admin_speed="auto",status="up",node="jamaica-01",op_duplex="full",op_flowcontrol="full",port="e0M",admin_duplex="auto"} 1500
net_port_status{datacenter="MUCCBC",cluster="jamaica",ignore_health="false",ipspace="Default",admin_flowcontrol="full",op_auto_negotiate="true",op_speed="1000",broadcast="Cluster-MGMT_v59",admin_auto_negotiate="true",admin_up="true",admin_speed="auto",status="up",node="jamaica-01",op_duplex="full",op_flowcontrol="full",port="e0M",admin_duplex="auto"} 0
net_port_port_mtu{datacenter="MUCCBC",cluster="jamaica",admin_flowcontrol="full",ignore_health="false",ipspace="Default",op_auto_negotiate="false",op_duplex="half",admin_duplex="auto",admin_speed="auto",admin_auto_negotiate="true",admin_up="true",status="down",node="jamaica-01",port="e11d"} 1500
net_port_port_admin_mtu{datacenter="MUCCBC",cluster="jamaica",admin_flowcontrol="full",ignore_health="false",ipspace="Default",op_auto_negotiate="false",op_duplex="half",admin_duplex="auto",admin_speed="auto",admin_auto_negotiate="true",admin_up="true",status="down",node="jamaica-01",port="e11d"} 1500
net_port_status{datacenter="MUCCBC",cluster="jamaica",admin_flowcontrol="full",ignore_health="false",ipspace="Default",op_auto_negotiate="false",op_duplex="half",admin_duplex="auto",admin_speed="auto",admin_auto_negotiate="true",admin_up="true",status="down",node="jamaica-01",port="e11d"} 0

[jgasher]

Nice! I’ll try it later today.

[ruanruijuan]

@jgasher As we are migrating from ZAPI collection to RESTAPI (issue #59), the following attributes from the net-port-get-iter ZAPI do not have RESTAPI equivalent. We have gone through the RESTAPI gap analysis with various ONTAP teams for Unified Manager and was told that 1) ONTAP has done most of the RESTAPI development and not every ZAPI/ZAPI attribute would have a RESTAPI equivalent; 2) we would need to give justifications why certain RESTAPI gaps need to be closed; 3) ONTAP has one last release of 9.11 to close the remaining gaps (if they agree with the gaps).

administrative-duplex
administrative-flowcontrol
administrative-speed - it's being deprecating in CLI
ignore-health-status
is-administrative-auto-negotiate
is-operational-auto-negotiate
mtu-admin
operational-duplex - default is always full-duplex
operational-flowcontrol

How important for you (or your respective customers) to include these attributes? If we add them now into Harvest, chances are there will be no data for them once we switch over to RESTAPI.

[jgasher]

I don’t have a business justification. It is fine to not have these in the future.

[rahulguptajss]

Rest Collector has netport template enabled by default while Zapi collector has it disabled. It should be same.

[rahulguptajss]

verified in 22.11