Using the Path Visualization View

Last updated: Wed Apr 03 21:15:08 GMT 2019

When you create any test which includes network metrics – including Network, HTTP Server (with network measurements enabled), Page load (with network measurements enabled), or DNS server (with network measurements enabled), or Voice, you get access to the Path Visualization view for the test.

Interface components

The screenshot below shows the Path Visualization view for an Agent to Agent Test.
Path_Visualization_Fowarding_Loss_update.png

  1. Test selector – use this control to change which test is being displayed in the view.  All tests will be shown here, including disabled tests and saved events.  
  2. Views available for this test – shows a list of shortcuts to other views for the selected test.
  3. Timeline – shows the lesser of 30 days, or the length of time that the test has been enabled, by charting the average values for the metric selected (see #6).  To change the data displayed, either use the 24h (trailing 24 hours) / 7d (trailing 7 days) / 14d (trailing 14 days) links, click the zoom out and latest buttons, or click and highlight the requested area of the chart, and it will re-draw appropriately.  Using the lower timeline will allow you to select a specific date or time and zoom in on events.
  4. Agents selector – as found elsewhere in ThousandEyes views, this controls which agent is charted on the timeline.  When an agent is selected in this control, it will be automatically selected in the timeline.
  5. Metric selector– choose between the various network metrics available; these metrics are listed below.  Changing the selected metric will change the timeline (#3, below) to reflect that metric, and change the coloring of agents in the visualization:
    • Loss: a timeline with a flat set of results (showing 0% loss) is important for stability.  Agents are colored on a scale from green to red, with green indicating 0% end-to-end loss, and red indicating 100% end-to-end loss.   
    • Latency: a timeline with a flat set of results is an indication of a network with low latency.  Spikes indicate an increase in the amount of latency in the connection, and should be investigated if they are sustained.  Agents are colored on a scale from green to red, with green indicating low latency, and red indicating higher latency.
    • Jitter: A timeline with a flat set of results is an indication of a very stable network.  Increases in jitter are bad for networks which are used for streaming, and indicate a variation in latency/response time.  Spikes in jitter are usually correlated to increases/dips in latency.
    • Bandwidtha timeline with flat set of results indicative of stable bandwidth availability. Spikes in the bandwidth graph indicate increase in network traffic load and are bad for streaming networks and bandwidth intensive applications.
    • Mean Opinion Score (MOS): a timeline with a flat set of results indicates stable quality voice call. Agents are colored on a scale from green to red, with green indicating MOS of 5, and red indicating MOS of 1.
    • Discards: a timeline with a flat set of results (showing 0% discards) is important for continuous voice playback. Discard spikes indicate intermittent pauses in voice playback. Agents are colored on a scale from green to red, with green indicating 0% discards, and red indicating 100% discards.  
    • Packet Delay Variation (PDV): a flat timeline measures least variation in packets and indicates stable voice call performance. Agents are colored on a scale from green to red, with green indicating least PDV (0 ms), and red indicating max PDV (40 ms).
    Refer to Knowledge Base article https://support.thousandeyes.com/ViewArticle?articleIdParam=kA0E0000000CmmzKAC for more information on metrics captured by the platform. By default, the loss metric is selected while for Voice tests, the default metric selected is MOS.
  6. Date/time label – when an instant in time is selected on the timeline, shown by the grey window on the upper timeline, the time and date of the selection are displayed here.  By default, the date/time selected in the previous view is selected, but if you enter the page directly from the views menu, the latest data will be selected by default.
  7. Quick selection – not always visible, the quick selection dropdown will highlight areas of interest on the visualization, based on the metric selected (#6 above).  The quick selection area will show the following type of links:
    • X links are part of an MPLS tunnel: this indicates that an MPLS network was found in the path from an agent to the destination.  When an MPLS link is found and selected in the visualization, the MPLS label is shown in the information box, similar to the figure to the right of this text, which can assist in diagnosis of an MPLS routing problem within an internal network.
      Screen_Shot_2015-07-08_at_12.23.44_AM.png
    • X links with MTU < 1500 bytes: this helps to detect the presence of different network types, in the path between the source and destination: http://en.wikipedia.org/wiki/Maximum_transmission_unit
    • X links with delay > Y ms: the presence of this link is controlled by the link markup control (#9).  Moving the delay slider controls the number of links shown and the amount of delay met by this link.
    • X nodes with loss > Y %: the presence of this link is controlled by the loss slider on the link markup control (#9).  Moving the loss slider controls the number of nodes shown and the loss percentage met by this link.  If no nodes show any loss, then this link will not be displayed.
    • X agents with mismatched DSCP: this link is visible when the received DSCP value is different from the one sent in the probe.
      Screen_Shot_2014-08-20_at_2.20.28_PM.png

8. Agent and Endpoint complexity controls: This control is also present in the path visualization view, but it effectively allows simplification of the Path Visualization by collapsing routes between various nodes on the Path Visualization.  The further apart the controls on the slider are, the less complex the visualization.
 
Screen_Shot_2015-07-08_at_12.24.12_AM.png

9. Highlight controls - allows the user to quickly identify nodes in a specific network (Autonomous System), country, or based on IP address, rDNS entry, or network Prefix. To use this control, enter some text into the search box and select one or more entries from the resulting list.When this option is used, nodes in the visualization will be highlighted (see highlighted node in legend below). 

10. Link/Node controls - these controls modify the content of the Quick Selection section (#11), and change the color of nodes and links in the path visualization:

Screen_Shot_2016-06-21_at_5.00.18_PM.png

  • Color links with delay > X ms - this option defaults to 100ms, and is based on a sliding scale, beginning at 10ms.  When you change the value shown here, it changes how links are colored in the visualization.  Sliding the delay lower will increase the number of red links (see link with delay in legend below) shown in the visualization.
  • Mark nodes with forwarding loss > X% - this option defaults to 10%, and runs from 0 to 100%.  If any node in the visualization exhibits loss greater than or equal to the value shown in this control, the node will be outlined with a red ring (see node with loss in legend below).  This will also control the presence of X nodes with loss > Y% link in the Quick Selection section.

11. Selecting dropdown - when items are selected, the object details section shows a link to expand the object details.  When you hover over an item in the selected objects table, its size doubles, to allow the user to identify the object in the visualization.  Click the link again to hide the table view of selected objects.  

Path_vis_forwarding_loss_2.jpeg

12. Group interfaces - found above the Agent and Endpoint allows multiple IP addresses to be grouped together.  With managed interfaces, the concept is very simple: create a grouping of IP addresses which are represented as a single node on a path visualization.  This would allow us to converge three IP addresses into a single element on the path visualization. 

ungrouped-interfaces.png

When multiple interfaces are grouped together, they appear as a slightly larger version of a node in path visualization, and allow a simplified path visualization view, similar to the image below.

grouped-interfaces.png

The interface grouping management can be turned on or off, by using the group interfaces feature, which is now checked by default.  This can be found above the origin and endpoint complexity controls on the path visualization page.  Clicking the manage groups link, allows you to create interface groups, which allow simplification of the path visualizations.  The following image shows an example of the interface groups management; in this example, the addresses 1.1.1.2 and 1.1.2.2 are grouped into a single interface group.

Interface_Groups_Management.png

13. Node Names - Simply check this box to show node labels (which will show the IP address or interface group name) for the node. If a node name is too long to display without obscuring the visualization, the label will be truncated - on a middle-out basis.


Interacting with the Path Visualization view

Hover over any agent/node/link on the path visualization to pop up detailed information. Agents show the name, IP address, prefix, network, location and destination IP, as well as metrics statistics (#6 above). Hovering on the nodes gives information on IP address, prefix, network, location and average response time. Links show their respective source and destination IP address, number of routes using that link, MPLS information and average link delay where possible.

Click a node or a link to select it.  Multiple objects can be selected at once.  Once an object is selected, it will show a moving dashed line surrounding the object.  Entire paths can be selected by double-clicking a node or a link.  All objects which traverse the object which is double clicked will have selection toggled on or off, based on the selection status of the object when it is selected.  

Once one or more objects is selected, the selected objects table link is activated. The link may display anything in the format (X nodes) (and) (Y links) selected.  Expand a table view of selected objects.  This will show two tabs: one for nodes, and one for links.

  • nodes shows Node Name, AS number, and an X button to de-select the node.
  • links shows the source and destination of each link, the number of routes (shown in the visualization) which transit that link, and the average link delay.

All nodes can be repositioned by clicking and dragging the node to a new place on the visualization pane, but the repositioned placement does not persist when the visualization is redrawn due to a change of time/date or display options.

 

Legend

Object ImageWhat it signifiesComments
DRAFT-pathvis-agent-green.pngDRAFT-pathvis-agent-orange.pngDRAFT-pathvis-agent-red.pngAgent locationAgent changes color based on the metric selected.
Scale goes from dark green (no loss, latency, jitter, etc.) to red (severe loss, latency, jitter, etc.)
DRAFT-privateagent.pngEnterprise AgentThe color of a Enterprise Agent is double-ringed, and changes color according to the same scale as a ThousandEyes Cloud Agent.
DRAFT-pathvis-bluedot.pngIdentifiable nodeA blue node indicates that IP information is available. 
DRAFT-pathvis-whitedot.pngUnidentifiable node A white node indicates that IP information is not available.
blue_dot.pngNode in local networkA dark blue node indicates that a node was identified inside the agent's source network.
DRAFT-pathvis-greendot.pngNode in destination network A node shaded in green indicates a node which was identified as inside the destination network, as specified by the Autonomous System of the customer.
DRAFT-pathvis-highlightednode.pngHighlighted node A node can be highlighted using the country/network selector. This helps to quickly identify nodes in a specific country or network.
DRAFT-pathvis-terminalnode.pngNode with lossA node circled in red indicates that loss is occurring at that point in the path, meeting the percentage threshold specified by the loss slider.
DRAFT-pathvis-greendot-terminal.pngEndpoint nodeA node circled in black, and showing an IP address beside it is an endpoint (or target) of the test. 
DRAFT-pathvis-selectedagent.pngSelected nodeA node circled in a moving blue dashed line indicates that the node is selected.
DRAFT-pathvis-selectedlink.pngSelected linkA link represented as a moving blue dashed line indicates that the link is selected.
 DRAFT-pathvis-dottedpath.pngCollapsed pathA path showing a dotted line is an indication of a path which was simplified for visualization purposes.  Expand using the complexity slider, or by clicking the label indicating the number of hops which were collapsed. 
 DRAFT-pathvis-splitpath.pngSplit pathA path showing a split is an indication that there are multiple routes to the destination.  All path visualization is based on a minimum of three tests running from each agent.  
When a path splits, the thickness of the line representing the link between the nodes will show how many of the tests traversed each link.
 DRAFT-pathvis-linkwithdelay.pngLink with high delayA red link will indicate a delay meeting the threshold specified by the latency slider.
Dotted line with a question markUnknown number
of hops between nodes
A dotted link with a question mark indicates insufficient data to determine the number of hops separating these nodes. Typically indicative of differing numbers of unresponsive nodes (* characters) between responsive nodes, or an indication of path trace being unable to reach the destination when the end-to-end measurement was performed successfully.
Dotted line with an X characterUnable to reach target nodeA dotted link with an X symbol indicates a trace that was unable to be completed to the target due to 100% forwarding and 100% end-to-end loss.