If you are using the current version of Cumulus NetQ, the content on this page may not be up to date. The current version of the documentation is available here. If you are redirected to the main page of the user guide, then this page may have been renamed; please search for it there.

Manage NetQ Agents

Run the following commands to view the status of an agent, disable an agent, manage logging, and configure the events the agent collects.

View NetQ Agent Status

The syntax for the NetQ Agent status command is:

netq [<hostname>] show agents
    [fresh | dead | rotten | opta]
    [around <text-time>]
    [json]

You can view the status for a given switch, host or NetQ Appliance or Virtual Machine. You can also filter by the status and view the status at a time in the past.

To view the current status of all NetQ Agents, run:

cumulus@switch~:$ netq show agents

To view NetQ Agents that are not communicating, run:

cumulus@switch~:$ netq show agents rotten
No matching agents records found

To view NetQ Agent status on the NetQ appliance or VM, run:

cumulus@switch~:$ netq show agents opta
Matching agents records:
Hostname          Status           NTP Sync Version                              Sys Uptime                Agent Uptime              Reinitialize Time          Last Changed
----------------- ---------------- -------- ------------------------------------ ------------------------- ------------------------- -------------------------- -------------------------
netq-ts           Fresh            yes      3.2.0-ub18.04u30~1601393774.104fb9e  Mon Sep 21 16:46:53 2020  Tue Sep 29 21:13:07 2020  Tue Sep 29 21:13:07 2020   Thu Oct  1 16:29:51 2020

View NetQ Agent Configuration

You can view the current configuration of a NetQ Agent to determine what data it collects and where it sends that data. The syntax for this command is:

netq config show agent 
    [cpu-limit|frr-monitor|kubernetes-monitor|loglevel|sensors|ssl|stats||wjh|wjh-threshold] 
    [json]

The following example shows a NetQ Agent in an on-premises deployment, talking to an appliance or VM at 127.0.0.1 using the default ports and VRF. There is no special configuration to monitor Kubernetes, FRR, interface statistics, sensors, or WJH, and there are no limits on CPU usage or change to the default logging level.

cumulus@switch:~$ netq config show agent
netq-agent             value      default
---------------------  ---------  ---------
exhibitport
exhibiturl
server                 127.0.0.1  127.0.0.1
cpu-limit              100        100
agenturl
enable-opta-discovery  True       True
agentport              8981       8981
port                   31980      31980
vrf                    default    default
()

To view the configuration of a particular aspect of a NetQ Agent, use the various options.

This example shows a NetQ Agent configured with a CPU limit of 60%.

cumulus@switch:~$ netq config show agent cpu-limit
CPU Quota
-----------
60%
()

Modify the Configuration of the NetQ Agent on a Node

The agent configuration commands let you:

  • Add, disable, and remove a NetQ Agent
  • Start and stop a NetQ Agent
  • Configure a NetQ Agent to collect selected data (CPU usage limit, FRR, Kubernetes, sensors, WJH)
  • Configure a NetQ Agent to send data to a server cluster
  • Troubleshoot the NetQ Agent

Commands apply to one agent at a time, and you run them on the switch or host where the NetQ Agent resides.

Add and Remove a NetQ Agent

To add or remove a NetQ Agent, you must add or remove the IP address (and port and VRF when specified) from the NetQ configuration file (at /etc/netq/netq.yml). This adds or removes the information about the appliance or VM where the agent sends the data it collects.

To use the NetQ CLI to add or remove a NetQ Agent on a switch or host, run:

netq config add agent server <text-opta-ip> [port <text-opta-port>] [vrf <text-vrf-name>]
netq config del agent server

If you want to use a specific port on the appliance or VM, use the port option. If you want the data sent over a particular virtual route interface, use the vrf option.

This example shows how to add a NetQ Agent and tell it to send the data it collects to the NetQ Appliance or VM at the IPv4 address of 10.0.0.23 using the default port (on-premises = 31980; cloud = 443) and vrf (default).

cumulus@switch~:$ netq config add agent server 10.0.0.23
cumulus@switch~:$ netq config restart agent

Disable and Reenable a NetQ Agent

You can temporarily disable the NetQ Agent on a node. Disabling the NetQ Agent maintains the data already collected in the NetQ database, but stops the NetQ Agent from collecting new data until you reenable it.

To disable a NetQ Agent, run:

cumulus@switch:~$ netq config stop agent

To reenable a NetQ Agent, run:

cumulus@switch:~$ netq config restart agent

Configure a NetQ Agent to Limit Switch CPU Usage

While not typically an issue, you can restrict the NetQ Agent from using more than a configurable amount of the CPU resources. This setting requires Cumulus Linux versions 3.6.x, 3.7.x or 4.1.0 or later to be running on the switch.

For more detail about this feature, refer to this Knowledge Base article.

This example limits a NetQ Agent from consuming more than 40% of the CPU resources on a Cumulus Linux switch.

cumulus@switch:~$ netq config add agent cpu-limit 40
cumulus@switch:~$ netq config restart agent

To remove the limit, run:

cumulus@switch:~$ netq config del agent cpu-limit
cumulus@switch:~$ netq config restart agent

Configure a NetQ Agent to Collect Data from Selected Services

You can enable and disable data collection about FRR (FRRouting), Kubernetes, sensors, and WJH (What Just Happened).

To configure the agent to start or stop collecting FRR data, run:

cumulus@chassis~:$ netq config add agent frr-monitor
cumulus@switch:~$ netq config restart agent

cumulus@chassis~:$ netq config del agent frr-monitor
cumulus@switch:~$ netq config restart agent

To configure the agent to start or stop collecting Kubernetes data, run:

cumulus@switch:~$ netq config add agent kubernetes-monitor
cumulus@switch:~$ netq config restart agent

cumulus@switch:~$ netq config del agent kubernetes-monitor
cumulus@switch:~$ netq config restart agent

To configure the agent to start or stop collecting chassis sensor data, run:

cumulus@chassis~:$ netq config add agent sensors
cumulus@switch:~$ netq config restart agent

cumulus@chassis~:$ netq config del agent sensors
cumulus@switch:~$ netq config restart agent

This command is only valid when run on a chassis, not a switch.

To configure the agent to start or stop collecting WJH data, run:

cumulus@chassis~:$ netq config add agent wjh
cumulus@switch:~$ netq config restart agent

cumulus@chassis~:$ netq config del agent wjh
cumulus@switch:~$ netq config restart agent

Configure a NetQ Agent to Send Data to a Server Cluster

If you have a server cluster arrangement for NetQ, you should configure the NetQ Agent to send the data it collects to every server in the cluster.

To configure the agent to send data to the servers in your cluster, run:

netq config add agent cluster-servers <text-opta-ip-list> [port <text-opta-port>] [vrf <text-vrf-name>]

You must separate the list of IP addresses by commas (not spaces). You can optionally specify a port or VRF.

This example configures the NetQ Agent on a switch to send the data to three servers located at 10.0.0.21, 10.0.0.22, and 10.0.0.23 using the rocket VRF.

cumulus@switch:~$ netq config add agent cluster-servers 10.0.0.21,10.0.0.22,10.0.0.23 vrf rocket

To stop a NetQ Agent from sending data to a server cluster, run:

cumulus@switch:~$ netq config del agent cluster-servers

Configure Logging to Troubleshoot a NetQ Agent

The logging level used for a NetQ Agent determines what types of events get logged about the NetQ Agent on the switch or host.

First, you need to decide what level of logging you want to configure. You can configure the logging level to be the same for every NetQ Agent, or selectively increase or decrease the logging level for a NetQ Agent on a problematic node.

Logging LevelDescription
debugSends notifications for all debug, info, warning, and error messages.
infoSends notifications for info, warning, and error messages (default).
warningSends notifications for warning and error messages.
errorSends notifications for errors messages.

You can view the NetQ Agent log directly. Messages have the following structure:

<timestamp> <node> <service>[PID]: <level>: <message>

ElementDescription
timestampDate and time event occurred in UTC format
nodeHostname of network node where event occurred
service [PID]Service and Process IDentifier that generated the event
levelLogging level assigned for the given event: debug, error, info, or warning
messageText description of event, including the node where the event occurred

For example:

logging message anatomy, including timestamp, node, service, level, and message

To configure a logging level, follow these steps. This example sets the logging level to debug:

  1. Set the logging level:

    cumulus@switch:~$ netq config add agent loglevel debug
    
  2. Restart the NetQ Agent:

    cumulus@switch:~$ netq config restart agent
    
  3. (Optional) Verify connection to the NetQ appliance or VM by viewing the netq-agent.log messages.

Disable Agent Logging

If you set the logging level to debug for troubleshooting, NVIDIA recommends that you either change the logging level to a less verbose mode or disable agent logging when you finish troubleshooting.

To change the logging level from debug to another level, run:

cumulus@switch:~$ netq config add agent loglevel [info|warning|error]
cumulus@switch:~$ netq config restart agent

To disable all logging:

cumulus@switch:~$ netq config del agent loglevel
cumulus@switch:~$ netq config restart agent

Change NetQ Agent Polling Data and Frequency

The NetQ Agent contains a pre-configured set of modular commands that run periodically and send event and resource data to the NetQ appliance or VM. You can fine tune which events the agent can poll and vary frequency of polling using the NetQ CLI.

For example, if your network is not running OSPF, you can disable the command that polls for OSPF events. Or you can decrease the polling interval for LLDP from the default of 60 seconds to 120 seconds. By not polling for selected data or polling less frequently, you can reduce switch CPU usage by the NetQ Agent.

Depending on the switch platform, the NetQ Agent might not execute some supported protocol commands. For example, if a switch has no VXLAN capability, then the agent skips all VXLAN-related commands.

Supported Commands

To see the list of supported modular commands, run:

cumulus@switch:~$ netq config show agent commands
 Service Key               Period  Active       Command
-----------------------  --------  --------  ---------------------------------------------------------------------
bgp-neighbors                  60  yes       ['/usr/bin/vtysh', '-c', 'show ip bgp vrf all neighbors json']
evpn-vni                       60  yes       ['/usr/bin/vtysh', '-c', 'show bgp l2vpn evpn vni json']
lldp-json                     120  yes       /usr/sbin/lldpctl -f json
clagctl-json                   60  yes       /usr/bin/clagctl -j
dpkg-query                  21600  yes       dpkg-query --show -f ${Package},${Version},${Status}\n
ptmctl-json                   120  yes       ptmctl
mstpctl-bridge-json            60  yes       /sbin/mstpctl showall json
ports                        3600  yes       Netq Predefined Command
proc-net-dev                   30  yes       Netq Predefined Command
agent_stats                   300  yes       Netq Predefined Command
agent_util_stats               30  yes       Netq Predefined Command
tcam-resource-json            120  yes       /usr/cumulus/bin/cl-resource-query -j
btrfs-json                   1800  yes       /sbin/btrfs fi usage -b /
config-mon-json               120  yes       Netq Predefined Command
running-config-mon-json        30  yes       Netq Predefined Command
cl-support-json               180  yes       Netq Predefined Command
resource-util-json            120  yes       findmnt / -n -o FS-OPTIONS
smonctl-json                   30  yes       /usr/sbin/smonctl -j
sensors-json                   30  yes       sensors -u
ssd-util-json               86400  yes       sudo /usr/sbin/smartctl -a /dev/sda
ospf-neighbor-json             60  yes       ['/usr/bin/vtysh', '-c', 'show ip ospf vrf all neighbor detail json']
ospf-interface-json            60  yes       ['/usr/bin/vtysh', '-c', 'show ip ospf vrf all interface json']

The NetQ predefined commands include:

  • agent_stats: Collects statistics about the NetQ Agent every 5 minutes.
  • agent_util_stats: Collects switch CPU and memory utilization by the NetQ Agent every 30 seconds.
  • cl-support-json: Polls the switch every 3 minutes to determine if an agent generated a cl-support file.
  • config-mon-json: Polls the /etc/network/interfaces, /etc/frr/frr.conf, /etc/lldpd.d/README.conf, and /etc/ptm.d/topology.dot files every 2 minutes to determine if the contents of any of these files has changed. If a change occurred, the agent transmits the contents of the file and its modification time to the NetQ appliance or VM.
  • ports: Polls for optics plugged into the switch every hour.
  • proc-net-dev: Polls for network statistics on the switch every 30 seconds.
  • running-config-mon-json: Polls the clagctl parameters every 30 seconds and sends a diff of any changes to the NetQ appliance or VM.

Modify the Polling Frequency

You can change the polling frequency (in seconds) of a modular command. For example, to change the polling frequency of the lldp-json command to 60 seconds from its default of 120 seconds, run:

cumulus@switch:~$ netq config add agent command service-key lldp-json poll-period 60
Successfully added/modified Command service lldpd command /usr/sbin/lldpctl -f json

cumulus@switch:~$ netq config show agent commands
 Service Key               Period  Active       Command
-----------------------  --------  --------  ---------------------------------------------------------------------
bgp-neighbors                  60  yes       ['/usr/bin/vtysh', '-c', 'show ip bgp vrf all neighbors json']
evpn-vni                       60  yes       ['/usr/bin/vtysh', '-c', 'show bgp l2vpn evpn vni json']
lldp-json                      60  yes       /usr/sbin/lldpctl -f json
clagctl-json                   60  yes       /usr/bin/clagctl -j
dpkg-query                  21600  yes       dpkg-query --show -f ${Package},${Version},${Status}\n
ptmctl-json                   120  yes       /usr/bin/ptmctl -d -j
mstpctl-bridge-json            60  yes       /sbin/mstpctl showall json
ports                        3600  yes       Netq Predefined Command
proc-net-dev                   30  yes       Netq Predefined Command
agent_stats                   300  yes       Netq Predefined Command
agent_util_stats               30  yes       Netq Predefined Command
tcam-resource-json            120  yes       /usr/cumulus/bin/cl-resource-query -j
btrfs-json                   1800  yes       /sbin/btrfs fi usage -b /
config-mon-json               120  yes       Netq Predefined Command
running-config-mon-json        30  yes       Netq Predefined Command
cl-support-json               180  yes       Netq Predefined Command
resource-util-json            120  yes       findmnt / -n -o FS-OPTIONS
smonctl-json                   30  yes       /usr/sbin/smonctl -j
sensors-json                   30  yes       sensors -u
ssd-util-json               86400  yes       sudo /usr/sbin/smartctl -a /dev/sda
ospf-neighbor-json             60  no        ['/usr/bin/vtysh', '-c', 'show ip ospf vrf all neighbor detail json']
ospf-interface-json            60  no        ['/usr/bin/vtysh', '-c', 'show ip ospf vrf all interface json']

Disable a Command

You can disable unnecessary commands. This can help reduce the compute resources the NetQ Agent consumes on the switch. For example, if your network does not run OSPF, you can disable the two OSPF commands:

cumulus@switch:~$ netq config add agent command service-key ospf-neighbor-json enable False
Command Service ospf-neighbor-json is disabled

cumulus@switch:~$ netq config show agent commands
 Service Key               Period  Active       Command
-----------------------  --------  --------  ---------------------------------------------------------------------
bgp-neighbors                  60  yes       ['/usr/bin/vtysh', '-c', 'show ip bgp vrf all neighbors json']
evpn-vni                       60  yes       ['/usr/bin/vtysh', '-c', 'show bgp l2vpn evpn vni json']
lldp-json                      60  yes       /usr/sbin/lldpctl -f json
clagctl-json                   60  yes       /usr/bin/clagctl -j
dpkg-query                  21600  yes       dpkg-query --show -f ${Package},${Version},${Status}\n
ptmctl-json                   120  yes       /usr/bin/ptmctl -d -j
mstpctl-bridge-json            60  yes       /sbin/mstpctl showall json
ports                        3600  yes       Netq Predefined Command
proc-net-dev                   30  yes       Netq Predefined Command
agent_stats                   300  yes       Netq Predefined Command
agent_util_stats               30  yes       Netq Predefined Command
tcam-resource-json            120  yes       /usr/cumulus/bin/cl-resource-query -j
btrfs-json                   1800  yes       /sbin/btrfs fi usage -b /
config-mon-json               120  yes       Netq Predefined Command
running-config-mon-json        30  yes       Netq Predefined Command
cl-support-json               180  yes       Netq Predefined Command
resource-util-json            120  yes       findmnt / -n -o FS-OPTIONS
smonctl-json                   30  yes       /usr/sbin/smonctl -j
sensors-json                   30  yes       sensors -u
ssd-util-json               86400  yes       sudo /usr/sbin/smartctl -a /dev/sda
ospf-neighbor-json             60  no        ['/usr/bin/vtysh', '-c', 'show ip ospf vrf all neighbor detail json']
ospf-interface-json            60  no        ['/usr/bin/vtysh', '-c', 'show ip ospf vrf all interface json']

Reset to Default

To revert to the original command settings, run:

cumulus@switch:~$ netq config agent factory-reset commands
Netq Command factory reset successful