With NSX APIs, you can monitor both point-in-time values and real-time values of host transport node statistics.

The system uses the esx-obsrv-stats-monitor to monitor the health of ESXi host transport nodes.

Statistics are collected per host transport node. The values of the statistics are not aggregated at the level of each host cluster.

Point-in-time values of the host transport node statistics are collected from each host transport node periodically (default is every 300 seconds) and cached by the system. The last updated cached values are fetched from the host and returned by the API.

For debugging purposes, NSX users can monitor the real-time values of the host transport node statistics. System fetches the statistics on a real-time basis from the data path kernel modules that are running on the host transport nodes.

With NSX APIs, you can view the following types of statistics for host transport nodes:
  • Packet statistics (packet_stats)
  • Fast path system statistics (fast_path_sys_stats)
  • Platform packet statistics (platform_packet_stats)
  • Platform CPU usage statistics (platform_cpu_usage_stats)
  • Fast path lcore usage statistics (fast_path_lcore_usage_stats)

In the API response, the statistics are organized based on the datapath modules that are collecting the statistics from the host transport nodes. For a description of the statistics, see NSX Host Transport Node Statistics.

The procedure in this documentation explains the API workflow for monitoring the point-in-time values and real-time values of the host transport node statistics. The steps in the procedure mention only the API URIs. For a detailed information about the API schema, parameters, and sample API response or request payload, refer to the NSX API Guide.

To learn about using the NSX Central CLI or the ESXi host CLI for viewing the transport node statistics, refer to the NSX Command-Line Interface Reference.

Procedure

  1. Run the following API to view the descriptions of all the statistics in the esx-obsrv-stats-monitor: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sha/monitors/esx-obsrv-stats-monitor
  2. Run the following API to retrieve the list of profiles of the esx-obsrv-stats-monitor: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sha/monitors/esx-obsrv-stats-monitor/profiles

    The API response displays the default profile and a user-defined profile (if created) for this monitor. The default profile is created by the system and it is applied to the system-created default group. The default group contains all the ESXi host transport nodes in the system. The default profile is activated by default on all the host transport nodes.

    To view the configuration of the default profile, run the following API:

    GET https://<nsx-mgr>/policy/api/v1/infra/sha/monitors/esx-obsrv-stats-monitor/profiles/default-profile
  3. (Optional) Run the following API to edit the configuration of the default profile: 
    PATCH https://<nsx-mgr>/policy/api/v1/infra/sha/monitors/esx-obsrv-stats-monitor/profiles/default-profile

    Copy the GET API response of the default profile and paste it in a text editor. Edit the configuration values of the profile. Finally, submit the edited configuration in the request body of this PATCH API.

    Sample Scenario:

    Let us assume that you want to apply a user-defined profile to a specific set of hosts (say A and B), but apply the default profile to all the remaining hosts in the system. In this case, you must create a custom group, which contains these two host transport nodes (A and B). Groups with only host transport node members can be created using NSX API. The UI does not support this functionality. After you have created the desired group, use this group in the applied_to_group_paths property of the user-defined profile.

    Use the same PATCH API, as mentioned earlier, to create the user-defined profile.

    For example, the following API creates a user-defined profile with ID "profile-1":

    PATCH https://<nsx-mgr>/policy/api/v1/infra/sha/monitors/esx-obsrv-stats-monitor/profiles/profile-1

    When this API is successful, the user-defined profile (profile-1) overrides the default profile on host transport nodes A and B. For all the remaining host transport nodes in the system, the default profile is applied.

  4. (Optional) Run the following API to read the monitor status on a specific host transport node: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sha/monitors/esx-obsrv-stats-monitor/status?transport_node_id=<node-id>

    To get the node_id of host transport nodes, run the following API:

    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes
  5. Run the following API to view the point-in-time values of the host transport node statistics for a specific node: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<host-tn-id>/statistics/monitor

    To get the node_id of the host transport node, use the same API, as mentioned in the previous step.

    This API returns the last updated cached values of the health monitoring statistics for the specified host transport node. By default, only the packet statistics are returned for the host transport node.

    As mentioned earlier in this section, the API can return the following types of health monitoring statistics for a host transport node:
    • Packet statistics (packet_stats). This type is the default.
    • Fast path system statistics (fast_path_sys_stats)
    • Platform packet statistics (platform_packet_stats)
    • Platform CPU usage statistics (platform_cpu_usage_stats)
    • Fast path lcore usage statistics (fast_path_lcore_usage_stats)

    When the type query parameter is omitted in the API URI, the API returns only the packet statistics by default.

    To view the statistics for a specific type or multiple types together, use the type query parameter in the API URI.

    Example 1: The following API returns the point-in-time values of the host transport node statistics for a single type: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<host-tn-id>/statistics/monitor?type=platform_cpu_usage_stats

    Example 2: The following API returns the point-in-time values of the host transport node statistics for multiple types together:

    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<host-tn-id>/statistics/monitor?type=platform_cpu_usage_stats,fast_path_sys_stats,platform_packet_stats

    Ensure that the types are specified in a comma-separated list and without any space before and after the commas.

  6. Run the following API to view the real-time values of the transport node health statistics for a specific node: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<host-tn-id>/statistics/debug

    As mentioned earlier in step 5, when the type query parameter is omitted in the API URI, the API returns only the packet statistics by default.

    To view the statistics for a specific type or multiple types together, use the type query parameter in the API URI.

    Example 1: The following API returns the real-time values of the host transport node statistics for a single type: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<host-tn-id>/statistics/debug?type=platform_cpu_usage_stats
    Example 2: The following API returns the real-time values of the host transport node statistics for multiple types together: 
    GET https://<nsx-mgr>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<host-tn-id>/statistics/debug?type=platform_cpu_usage_stats,fast_path_sys_stats,platform_packet_stats