Starting with NSX-T Data Center 3.2, the application will support the collection and storage of data up to one year. With this feature, along with the point-in-time data, you can also view time series metrics, such as CPU usage memory, disk usage, packets per second, bytes per second, and packet drop rate, for edge nodes and gateways. You can view the metrics by invoking time series metrics APIs. A few of these metrics are also available on NSX Manager interface along with the point-in-time data.

Using time series metrics, you can monitor the trend in key performance indicators, detect anomalies, perform before and after analysis, and get the historical context which can help in troubleshooting. Based on your role, you can view metrics of only those objects for which you have the authorization.

You can use the metrics APIs to fetch the time series metrics. For information about using the APIs, see APIs to Fetch Time-Series Metrics.

On the NSX Manager interface, the time series metrics is displayed as charts with respective keys and time filter on last one hour, last 24 hours, last one week, last one month, and last one year. You can view the following information for edge nodes:
  • CPU cores allocated
  • Alarms
    • Edge node - Overall alarm count
    • CPU - Alarms for Datapath CPU and Service CPUs
    • Disk - Overall disk alarms and alarms for each partition
    • Memory - Overall memory alarms and alarms for each memory pool
  • Uptime
  • CPU
    • System Load - You can view current load average and average system load trend.
    • Datapath CPU - Number of datapath CPU cores and their usage details, which include the average usage of all cores and the highest usage among cores. You can view current utilization, highest utilization for a single core, and average usage trend.
    • Service CPU - Number of service CPU cores and their usage details, which include the average usage of all cores and highest usage among cores. You can view current utilization, highest utilization for a single core, and average usage trend.
  • Memory
    • System Memory - Utilization of RAM on edge node. You can view current utilization and average usage trend.
    • Datapath Memory - Includes heap memory, memory pool, and resident memory. You can view current usage, current pool usage, and average usage trend.
    • Memory Pools - List of all memory pools along with their description and usage values except for QAT memory pool (of Bare Metal Edge) whose usage is always around 100%. The memory pools are:
      Name Description
      jumbo_mbuf_pool Packet Pool for Jumbo Frame Used by Ipsec Crypto Device
      common_mbuf_pool Datapath Common Packet Pool
      sp_pktmbuf_pool Datapath Slowpath Packet Pool
      fw_mon_msg Stateful Service Sync Message Pool
      vxstt4_frag_q Vxstt Fragment Pool for Reassembly
      pfstatepl3 Stateful Service State Pool
      pffqdnippl Stateful Service FQDN to IP Map Pool
      pffqdnsyncpl Stateful Service FQDN SYNC Pool
      pffqdndnpl Stateful Service FQDN Internal Pool
      pfdnsdnpl Stateful Service FQDN Internal Pool
      pfpktpl3 Stateful Service Fragmented Packet Pool
      pfsyncmbufpl3 Stateful Service SYNC Pool
      pf_fp_rule_node Stateful Service Rule Node Pool
      pf_fp_root_rule_node Stateful Service Root Rule Node Pool
      pf_tb_root_rule_node Stateful Service Fastpath Root Table Node Pool
      pfa_intattr_pl3 Stateful Service Integer Attribute Pool
      pfa_attrconn_pl3 Stateful Service Attribute Connection Pool
      pfa_ctx_pl3 Stateful Service Context Pool
      pfa_key_ace_pl3 Stateful Service Integer Attribute Key Pool
      pfa_value_ace_pl3 Stateful Service Integer Attribute Value Pool
      lb_pkt_pl3 Load Balancer Temp Packet Cache Pool
  • Disk
    • Total disk usage for all ext4 disk partitions and the list of RAM disk and disk partitions. Also, the available free space of each partition. You can view overall GB used, read latency, write latency, current usage, average usage trend by partition, and average disk latency trend.
  • Transport node status
  • Network interface
    • To view statistics of any network interface, click the graphic icon. You can view cumulative statistics like total packets and dropped packets and time series metrics like network utilization trend and dropped packets trend.
  • NAT rules statistics

Prerequisites

Ensure that you have deployed NSX Application Platform. For more details about deploying NSX Application Platform, see the Deploying and Managing NSX Application Platform guide. The Metrics feature is enabled by default when you deploy NSX Application Platform.

If you view statistics without deploying NSX Application Platform or without enabling the Metrics feature, the system displays a message to deploy NSX Application Platform and to enable the Metrics feature.

Procedure

  1. To view metrics of an edge node, perform the following steps:
    1. With admin privileges, log in to NSX Manager.
    2. Select System > Fabric > Nodes.
    3. Click the Edge Transport Nodes tab.
    4. Click the edge node that you want to monitor.
    5. Click the Monitor tab.
    Usage information for CPU, memory, and disk is displayed, as well as the node status, network interfaces, and NAT rule statistics.
  2. To view gateway or interface metrics, perform the following steps:
    1. With admin privileges, log in to NSX Manager.
    2. Select Networking > Network Topology.
    3. Click the required gateway. The system displays a gateway details window.
    4. Under Statistics, select the linked gateway or the interface link for which you want to view the metrics.