Skip to content

Summary

Enter Infrastructure > Network > Summary to view the network performance overview of the current workspace.

On the Summary page, the system combines time-series trend charts and the network path data list to display network traffic and connection details between clients and servers across four infrastructure types: Hosts, Pods, Deployments, and Services.

  • Time-series charts: network metric trend charts over different time ranges, with support for switching between bytes sent/received, TCP latency/jitter, TCP retransmission count/connection count/close count

  • Network path list: includes client, server, bytes sent, bytes received, TCP latency, TCP retransmission count, and other information

Page Overview

Quick Filters

On the left side of the Summary page, you can quickly filter data by the following dimensions:

1. Basic Dimensions

  • Transport protocol: TCP / UDP
  • Transport direction: incoming / outgoing
  • Host: source or destination host name
  • PID: process ID
  • Process: process name

2. Client Dimensions

  • IP: client IP address
  • Port: client port
  • IP type: IPv4 / IPv6
  • Pod name: Kubernetes Pod identifier
  • Namespace: Kubernetes namespace
  • Deployment name: Deployment identifier
  • Service name: Service identifier

3. Server Dimensions

  • IP: server IP address
  • Port: server port
  • Domain name: server domain name
  • IP type: IPv4 / IPv6
  • Pod name: Kubernetes Pod identifier
  • Namespace: Kubernetes namespace
  • Deployment name: Deployment identifier
  • Service name: Service identifier

Transport Direction and Transport Protocol Combinations

Network performance monitoring based on TCP and UDP protocols is currently supported. Combined with incoming and outgoing, the following 6 combinations are available:

  • incoming + all protocols
  • incoming + TCP
  • incoming + UDP
  • outgoing + all protocols
  • outgoing + TCP
  • outgoing + UDP

Details Page

Click a network data entry in the list to enter the details page.

Network Analysis

Click the Network Analysis tab on the details page to view network performance metric trends and connection details between the client and server.

Metric Overview

Displays key network metric statistics within the current time range:

  • Bytes sent
  • Bytes received
  • TCP connection count
  • TCP retransmission count
  • TCP close count

Trend Charts

Provides multi-dimensional time-series trend charts, with support for switching between:

  • Bytes sent/received: network traffic trends
  • TCP latency/TCP jitter: network latency and jitter
  • TCP retransmission count/TCP connection count/TCP close count: connection quality metrics

Network Connection Analysis

The list below displays specific network connection data:

  • Search: supports keyword search and filtering
  • Aggregate by IP dimension: when enabled, aggregates statistics by IP address to improve query efficiency and avoid excessive data volume
  • Display settings: click the "Settings" button to flexibly configure the fields displayed in the list

The list displays key connection metrics such as client, server, bytes sent, bytes received, TCP latency, and TCP retransmission count.

Network Flow

Click the Network Flow tab on the details page to view detailed network connection packets between the selected hosts.

The top of the page displays the association between the currently selected host and server_name.

Toolbar Operations

  • Time widget: supports selecting a relative time range (such as the last 1 day) or an absolute time range
  • Search: automatically applies the current filter conditions from the details page; supports DQL syntax queries, allowing you to enter conditions such as host: or server_name: for precise filtering
  • Jump to Network Flow: click the button in the upper right corner to go to the Network Flow list, with the current filter conditions applied automatically
  • Display items: click the button on the right to customize which list fields are shown or hidden

Data List

Displays network flow metrics including time, client, server, transport direction, PID, host, network type, service, bytes sent, bytes received, TCP retransmission count, and TCP connection count.

Network Path

Click the Network Path tab on the details page to view the hop-by-hop network path and probe quality from the source (client) to the destination (server).

Metric Overview

Displays key path probe metrics within the current time range:

Metric Description
Path status Overall completion status of the probe

  • Destination reachable indicates that the hop-by-hop probe has reached the destination
  • Partial path indicates that only some intermediate nodes were obtained and arrival at the destination was not confirmed
  • Failed indicates an abnormal probe execution
    		•
    End-to-end latency The average round-trip time (RTT) from the probe source to the destination, with the minimum/maximum latency range also displayed. This value is calculated by an independent end-to-end probe and reflects the overall link quality
    End-to-end no-response ratio The proportion of end-to-end probes that did not receive a response from the destination. The actual number of probe samples sent/received in response is shown in parentheses (for example, 10 / 10). This ratio only represents the end-to-end probe results and is not equal to the real packet loss rate of any intermediate router
    Path hop count The number of intermediate nodes (TTL hops) traversed by the current path. If multiple probes exist, the average/minimum/maximum hop counts are displayed
    Probe count The number of Traceroute probe rounds performed in this record (for example, 3 probe rounds in total). Each probe round starts independently from TTL 1 and is used to observe path stability
    About Latency and No Response

    The page displays two latency concepts:

    • End-to-end latency: the latency shown in the metric card, reflecting the overall round-trip time from the probe source to the destination.
    • Per-hop RTT: the RTT of each node in the topology, reflecting the round-trip time from the probe source to that intermediate node. It is not the link latency between two adjacent hops.

    Path Topology

    By default, the hop-by-hop path topology of the most recent probe is displayed. Click the upper right corner to switch to the Path Change View to compare paths across multiple probes.

    1. Topology Elements:

    • Source node: the source that initiates the probe, displaying information such as the source process, service, IP address, and port

    • Intermediate nodes (TTL 1, TTL 2 ...): the intermediate network nodes that packets traverse. Each node displays:

      • Node name: the reverse DNS domain name is displayed first; if unavailable, the IP address is displayed; if that is also unavailable, the TTL n placeholder is displayed
      • Reachability status: whether the node responded to this probe
      • RTT: the round-trip time from the probe source to this node (in milliseconds)
      • Ownership information: for public IP addresses, the ASN organization name and cloud provider ownership may be displayed

    • Node connections: indicate the order of the probe path. The connections themselves do not carry link-level latency or packet loss data.

    2. Path Change View:

    When multiple probes exist, switching to the Path Change View allows vertical comparison of the results of each probe round. The system aligns nodes across rounds by TTL + IP address, helping you observe whether the route has changed (such as in ECMP multi-path scenarios).

    Historical Probes

    Below the topology, a continuous color band displays the distribution of all probe results for the path within the selected time range.

    Color Status Description
    Green Destination reachable The probe successfully reached the destination
    Orange Partial path Only some intermediate nodes were obtained
    Red Probe failed Abnormal probe execution
    Gray Unknown status No data or no probe

    Click any color block to refresh the metric cards, topology, and result area above with the probe data at that point in time. Switching the time widget at the top re-renders the historical probe area. Up to 100 historical records are displayed.

    Node Details

    Click the source node or any intermediate node in the topology to expand the detailed information for that node below:

    Field Description
    IP address The IP address of the node
    Port The source or destination port
    Host The name of the source host that discovered the path
    Source process The name of the process that generated the network traffic (dynamic discovery scenarios)
    Service The name of the service that generated the network traffic (dynamic discovery scenarios)
    Probe response The response status of this node in the current probe, such as Response received or No response
    Note
    1. If a TTL node shows No response, it only means that the node did not return a response to this probe. This may be caused by rate limiting, firewall policies, or anonymous routing on intermediate network devices. It cannot be directly concluded that the device is down or the link is interrupted;
    2. The topology does not display "link latency" or "link packet loss" between adjacent nodes. For path quality, refer to the end-to-end latency and end-to-end no-response ratio at the top;
    3. Some fields (such as network operator, cloud provider, domain name resolution, source process/service) depend on the collector version or platform data enrichment capabilities. If they are missing, this is normal and does not indicate a data anomaly.

    You can view hosts, Pods, or other related views associated with the current network data.