Manual Chapter : Monitoring TCP application service data

Applies To:

Show Versions Show Versions

BIG-IQ Centralized Management

  • 8.0.0
Manual Chapter

Monitoring TCP application service data

Monitoring TCP statistics

Network performance of client-to-application communication relies on the efficiency of latency times and packet volume sent over the TCP protocol. When examining TCP data over time, consider these metrics to evaluate the user experience with the application server.
The image displays the dashboard of an AS3 TCP application service. Using this dashboard, you monitor and mitigate issues with your TCP application service.
TCP analytics is available to all application service types (Legacy and SC) with a TCP analytics profile.
To evaluate the data for all your TCP applications, go to
Monitoring
DASHBOARDS
Local Traffic
TCP
.

Identify TCP application service issues

You can analyze the data of a TCP application service to establish traffic status, or to troubleshoot performance issues.
  1. Go to
    Applications
    APPLICATIONS
    .
    A list of all configured applications are displayed.
  2. Select the application name that has the TCP application service you would like to monitor.
  3. Select the name of the TCP application service.
    The TCP application service dashboard is displayed.
  4. Use the icons located at the center of the screen to evaluate traffic data in the charts at the bottom of the screen.
To make quick edits to your application service's configuration, select the
CONFIGURATION
tab at the center of the screen.

Evaluate client-side information from TCP data

You can evaluate data about your client-side connections using TCP analytics. When viewing TCP connection data, you can analyze detected client IP addresses, or traffic volume data from the client-side. There are two ways to view data: evaluate a specific application service; or evaluate TCP traffic to all your applications. Use the process that best suits your system's needs.
For more information about the charts, dimensions and metrics referenced in this task, see
TCP charts
and
TCP dimensions and metrics
.
  1. To view an application service's client-side data:
    1. Go to
      Applications
      APPLICATIONS
    2. Select the application name that hosts your TCP application service.
    3. Select the name of the TCP application service.
      The application service's dashboard is displayed. Ensure the
      ANALYTICS
      tab at the center of the screen is selected
    4. Click the icon
      All Types
      under CLIENT.
    5. Using the menu to the left of the chart, select
      Client Side Connections
      .
      The chart displays the client-side connection status over the selected time period.
    6. To view the client IPs requesting access to your application, expand the dimension
      Remote Host IP Address
      , located in the dimension pane to the right of the chart.
      You can select a specific IP to filter data displayed in the charts and other dimensions.
    7. To view general traffic data, from the client-side, select any of the charts located in the menu to the left.
  2. To view client-side data for all TCP application services:
    1. Go to
      Monitoring
      DASHBOARDS
      Local Traffic
      TCP
      .
    2. From the dimensions pane, to the right of the charts, expand
      Requests Side
      and select
      Client-Side
      .
      This action will filter all charts and dimensions on the screen to only display client-side request data.
    3. To view the client IPs requesting access to your applications, expand the dimension
      Remote Host IP Address
      .
      The list displays all client IP addresses detected in traffic over the selected time period. You can select one or more IP addresses to further filter the chart and remaining dimension data.

TCP Charts

The charts that reflect TCP data allow you to analyze current traffic data, and troubleshoot any issues that may occur over time. The parameters found in the charts described can indicate the wait time over the course of the client-server communication channel. To view charts for a TCP application service, go to
Applications
APPLICATIONS
<Application_Name>
<TCP_Application_Service>
.
You can view TCP data for all your traffic (to all managed applications) by going to
Monitoring
DASHBOARDS
Local Traffic
TCP
.
The image shows which selections display the charts detailed below. These charts display a combination of both request and server side data. Click
CLIENT
or
SERVERS
to view that part of the transaction with the BIG-IP system. Click
Environment
to display the status of the BIG-IP device.
Round Trip Time (RTT)
A measure of time it takes to send a packet and receive an ACK message from the target host. Increased RTT over time indicates limited network bandwidth.
Goodput
The number of useful bytes, delivered by the network, to a destination over time. Low goodput in ratio to overall throughput indicates network issues, such as increased retransmits.
Delay States
A traffic condition in which the TCP protocol limits packet flow to avoid network congestion. During periods of congestion, the TCP protocol applies a mitigation algorithm to manage traffic flow according to the root cause of congestion. For more information about the traffic conditions that prompts a delay state, see TCP Dimensions and Metrics.
Connections
The average number of new and closed TCP connections. Monitoring your TCP connections can indicate the level of activity across your network.
Packets
The volume of transmitted data over the network, including payload, and other required data based on the packet format. The outcome of these packets over time can indicate a number of network issues including: congestion, server performance/capacity limits, or an attack.

TCP Dimensions and Metrics

The following list defines the metrics gathered to monitor TCP network connections and the dimension aspects from which data is observed. The information provided is found in the dimensions pane tables with the screens that display TCP data. You use this information to filter the on-screen data by specific dimension objects. For example, if you would like to view data that pertains to specific virtual servers. It is important note that all metric data is displayed as a unit over the time period selected for the screen.

TCP Dimensions

The following defines the dimensions found in the dimensions pane that report TCP traffic analytics.
BIG-IP Service Clusters
The name of the logical grouping of multiple BIG-IP devices that have the same configuration. This dimension only displays transaction details if you have configured a service cluster.
BIG-IP Host Names
The name of each BIG-IP system that processed the monitored transactions.
BIG-IP Blade Names
The number of the BIG-IP Blade in a cluster. For a standalone BIG-IP, the value is 0.
Applications
The name of each application service grouping that reports TCP traffic data.
Application Services
The name of each TCP application service reporting transaction data.
Virtual Servers
The name of each virtual server that processes the TCP transactions.
Request Sides*
The TCP connection either between the BIG-IP system and the client or the server (depending on the chart in view). The recorded metric is a measure of the communication protocol between the remote host and the BIG-IP virtual server.
Continent
The geographic domain, based on continent code, that initiated the TCP connection with the BIG-IP virtual server.
Subnets
The subnet of the client IP address that initiated the TCP request that was processed by the BIG-IP system.
Countries and Regions
Either the location of the country or region domain, based on the country code, that initiated the TCP connection with the BIG-IP virtual server.
City and Post Codes
The location, based on IP location lookup, that initiated the TCP connection with the BIG-IP virtual server.
Remote Host IP Addresses
The IP addresses of either the client or servers that have initiated a TCP connection with your BIG-IP system.
Subscriber Type
The defined subscriber type based on a Policy Enforcement Manager (PEM) profile. If you do not have PEM subscriber data configured to your BIG-IP system, this dimension will not display detailed data.
Subscriber Name
The defined subscriber name based on a Policy Enforcement Manager (PEM) profile. If you do not have PEM subscriber data configured to your BIG-IP system, this dimension will not display detailed data.
Radio Types
The defined radio type based on a Policy Enforcement Manager (PEM) profile. If you do not have PEM radio data configured to your BIG-IP system, this dimension will not display detailed data.
Towers
The defined tower based on a Policy Enforcement Manager (PEM) profile. If you do not have PEM tower data configured to your BIG-IP system, this dimension will not display detailed data.
User Provided Keys
The defined user provided keys based on a Policy Enforcement Manager (PEM) profile. If you do not have PEM user provided keys configured to your BIG-IP system, this dimension will not display detailed data.

TCP Metrics

TCP metrics reflect the connection volume, connection duration, responses, and contents of the TCP connections to your managed BIG-IP systems. Metric sets categorize the metric data according to an aspect of the TCP connection. The table below defines the metric set and the kind of metric data collected.
Metric Set
Metric Set Description
Metric
Metric Description
Delay States
The time (in ms) spent in each delay state by all TCP connections.
3WHS
Delay in outgoing data delivery due to 3-way handshake that starts a TCP connection. This state is accrued only if round-trip-time of the SYN or SYN-ACK could be estimated.
RWND
Delay in outgoing data delivery due to limited receive-window, and the remote host’s flow-control forces the BIG-IP system into an idle state.
RETX
Delay in outgoing data delivery due to retransmission. TCP re-sends data and/or waiting for acknowledgment of those re-transmissions. This may indicate lossy links in the data path, or overly aggressive congestion control (for example, a profile with Slow Start disabled or improperly set Packet Loss Ignore settings). Activating rate-pace in the TCP profile may also help.
Wait for ACK
Delay in outgoing data delivery due to wait for ACK. The BIG-IP system has sent all available data and is awaiting an ACK from the remote host. If this state is prevalent, it could be a short connection, or the upper layers or the server are forcing TCP to frequently pause to accept new data
Closing
Delay in ACK of the FIN request. The BIG-IP system has received acknowledgment of all data, sent the FIN, and is awaiting acknowledgment of the FIN.
CWND
Delay in outgoing data delivery due to limited congestion window. The TCP congestion window is holding available data. This is usually a legitimate response to the bandwidth-delay product and congestion on the packet path. In some cases, it might be a poor response to non-congestion packet loss (fixable using the Packet Loss Ignore profile options) or inaccurate data in the congestion metrics cache
Nagle
Delay in outgoing data delivery due to Nagle's algorithm holding sub-MSS size packets. If the Nagle state shows up frequently, disable Nagle's algorithm in the TCP profile
Send Buffer
Delay in outgoing data delivery due to a full send buffer. Correctable by increasing the Send Buffer size in the TCP profile
App
Delay in outgoing data delivery to application layer. The BIG-IP system has successfully delivered all available data. There is a delay in either of the remote hosts, or in the layers above TCP on the BIG-IP system
Rate Pace
Delay in outgoing data delivery due to rate pacing. This has no impact on throughput, and no action is required
Packets Lost
Packets lost over the TCP connection between BIG-IP and the remote hosts.
Avg/s
The average number of packets per second lost between BIG-IP and the remote host.
Total
The number of packets lost between BIG-IP and the remote hosts.
Packets Received
Packets received over the TCP connection between BIG-IP and the remote hosts.
Avg/s
The average number of packets per second received by BIG-IP from the remote host.
Total
The number of packets received by BIG-IP from the remote hosts.
Packets Sent
The overall packets sent by BIG-IP to the remote hosts.
Avg/s
The average number of packets per second sent by BIG-IP to the remote hosts.
Total
The number of packets sent by BIG-IP to the remote host.
Goodput Received
The number of application-level bytes, or Goodput, received by BIG-IP from the remote hosts.
Avg/s
Average Goodput in bytes per second, received by BIG-IP from the remote hosts
Total
Total Goodput in bytes, received by BIG-IP from the remote hosts
Max
Highest value of Goodput observed, in bytes per second, for a single connection received by BIG-IP from a remote host
Min
Lowest value of Goodput observed, in bytes per second, for a single connection received by BIG-IP from the remote hosts
Goodput Sent
The number of application-level bytes, or Goodput, sent by BIG-IP to the remote hosts.
Avg/s
Average Goodput in bytes per second, sent from BIG-IPs to the remote hosts
Total
Total Goodput in bytes, sent from BIG-IP to the remote hosts
Max
Highest value of Goodput observed, in bytes per second, for a single connection sent from BIG-IP to remote hosts
Min
Lowest value of Goodput observed, in bytes per second, for a single connection sent from BIG-IP to remote hosts
Round Trip Time
The communication time (in ms) from a SYN to an ACK message between BIG-IP and the remote hosts
Avg
The average round trip time between BIG-IP and the remote hosts
Max
The highest observed round trip time between BIG-IP and the remote hosts
Min
The lowest observed round trip time between BIG-IP and the remote hosts.
Count
The number of times a round trip time was measured (number of data samples).
Connections Opened
Number of new TCP connections opened between BIG-IP and the remote hosts
Avg/s
The average number per second of TCP connections that were opened between BIG-IP and the remote hosts.
Total
The number of TCP connections that were opened between BIG-IP and the remote hosts
Connections Duration
Approximate time (in ms) TCP connections from the moment they are established until they are closed
Duration
The average duration of a TCP connection with BIG-IP from the moment they were established with the remote hosts, and until they were closed.
Connections Closed
The number of TCP connections between BIG-IP and remote hosts that were closed
Avg/s
The average number per second of TCP connections that were closed opened between BIG-IP and the remote hosts
Total
The number of TCP connections that were closed opened between BIG-IP and the remote hosts