Components: Topology
This workspace provides a visual representation of the connections between different components on your system. You can move the mouse over a node or link to display hover help, or right-click on a node to display a more detailed workspace for a particular component.
Using this workspace
The Topologies workspace displays a topology of interactions within your enterprise. The interactions are shown as links between nodes. The node appearance may change if Transaction Tracking can detect which component it refers to. Depending on which category you are in, the nodes in the Topologies workspace indicate a different level of detail:
- The Transactions category is the most detailed, and displays the interconnections between all transaction interactions that have occurred within a set time frame.
- The Applications category displays the interconnections between different applications.
- The Components category displays the interconnections between components.
- The Servers category displays the interconnections between servers.
Use the Topologies toolbar to change the view of the default topology. For example, you can zoom in on a particular node or link, or show the state of a node or link by selecting an icon from the toolbar. Move your mouse over items in the Topologies toolbar to display hover help about their function.
Using the Topologies workspace you can identify response times of specific nodes, and observe which nodes in your system interact with each other. The Topologies workspace contains two views in addition to the standard Navigator view:
- The topology, called Application Aggregate Topology, Component Aggregate Topology, Server Aggregate Topology, or Transaction Aggregate Topology depending on the category you select.
- A table called Applications, Components, Servers, or Transactions depending on the category you select.
The table below the topology contains the list of all aggregates. Click a link icon in the table to display more detailed workspaces related to that aggregate. If you link to a more detailed topology workspace, the node that you select will be the focal point of the new topology display, and the surrounding nodes with connecting links indicate how they interact with the selected node. You can identify each node by its name which is displayed in the topology.
To access further information about a node, move the mouse over a node in the topology to display hover help. The hover help displays the node category and name. In addition, if there is sufficient aggregate information, it displays the Average Time, Baseline and Deviation in milliseconds for the transaction of that node.
If the topology is well known, but there is not yet enough information to provide aggregate data, hover help will show basic information as soon as some data is available. An example of this is when you request aggregate data for a time period that has only just started.
If there is not enough data, the Average Time, Baseline, and Deviation will display 0.
When a transaction topology of a composite application changes, the current period will not reflect these changes in the topology. The time period before the change will be correct, and the time period following the current period will be correct. After the current period expires, the correct topology will be displayed. The default expiration of a period is five minutes.
The Applications, Components, Servers, or Transactions view provides a table with more details on individual interactions. Below is a description of the fields in this table.
| Column | Description |
|---|---|
| Name | The name of the Application, Component, Server, or Transaction. |
| Total Time | Measures the average time for a transaction
to complete, in milliseconds. The value in this column does not include
the time taken for failures, unless all transactions have failed.
The Total Time measures the following transactions
based on which category you are looking at:
|
| Total Time Deviation | Measures the percentage deviation from a baseline value. Positive numbers indicate a slower total time than the baseline value, and negative numbers indicate a faster total time than the baseline value. A deviation of zero indicates either that there is no deviation or that the deviation has not yet been measured. |
| Transaction Rate | Measures the average rate of transactions. This
is measured in transactions per minute. The Transaction
Rate calculates the following rate of transactions based
on which category you are looking at:
|
| Transaction Rate Deviation | Measures the percentage deviation from a baseline value in the number of transactions. Positive numbers indicate more transactions than the baseline value, and negative numbers indicate fewer transactions than the baseline value. A deviation of zero indicates either that there is no deviation or that the deviation has not yet been measured. |
| Percent Failed | Indicates the percentage of transaction instances that have failed. |
| Percent Slow | Indicates the percentage of transaction instances that are running slower than a preset threshold. |
| Percent Good | Indicates the percentage of transaction instances that are running faster than a preset threshold. |
| Timestamp | Indicates the start time of the period of aggregation. |
| Enclosing Component | Enclosing component aggregate if applicable. |
| Enclosing Server | Enclosing server aggregate if applicable. |
Interaction Rows at an Aggregate level are used to draw the topologies at the Applications, Components, Servers, and Transactions workspaces. The Transaction Reporter generates these rows from Aggregates retrieved from Transaction Collectors, and a topology that it determines from Instance Data. At startup the Transaction Reporter has no known topology. When the first Aggregates are collected from the Transaction Collectors, the Transaction Reporter will obtain some instance data and perform a multi-hop trace. The Transaction Reporter uses the results to determine that instances associated with one Aggregate Record interact with instances from another Aggregate Record. This is the basis of creating Interaction Rows at the Aggregate Level. If there are a large number of Transaction Collectors or a large number of interactions, the initial traces may take tens of minutes to perform. Since Transaction Collectors produce Aggregates every 5 minutes, the Transaction Collector has a default configuration to spend a maximum of 4 minutes performing these "one-hop" traces before producing Interaction Rows, and moving to the next Aggregation Period. This will result in incomplete Interaction Rows for those initial aggregation periods.
The following parameters relate to the creation of Aggregate Interaction Rows and the hop trace:
- RECALCULATE_INTERACTIONS=N
- CALCULATE_IMPLIED_INTERACTIONS=Y
- CALCULATE_PSEUDO_INTERACTIONS=Y
- MAXIMUM_PROCESS_AGGREGATES_MINUTES=4
- MAXIMUM_TOPOLOGY_TRACE_BACKOFF_MINUTES=120
- INITIAL_TOPOLOGY_TRACE_EVENT_COUNT=2
- MAXIMUM_TOPOLOGY_TRACE_EVENT_COUNT=15
- DETERMINE_TOPOLOGY_TRACE_HOP_COUNT=3
- TOPOLOGY_FORGET_INTERVAL_MINUTES=0
The Transaction Reporter generates Interaction Rows at an Instance level directly from Instance data obtained from Transaction Collectors. An initial instance from which to perform the "full" instance trace is obtained from a Transaction Collector as one that is associated with an Aggregate. The Transaction Collector provides the Transaction Reporter with the identities of the five average, five fastest and five slowest transactions that were used to generate a particular aggregate. Select one of these via the Instances workspace. Afterwards, when selecting an Instance Topology, the Transaction Reporter requests a "full" trace. If there are a large number of Transaction Collectors or a large number of interactions, the full trace may take some minutes to perform. In this case, the workspace may time out without receiving any instance interaction rows. The Transaction Reporter has a number of parameters in the ENV file to limit the time or depth of a full trace. If these parameters are set, the instance interaction rows determined to that point will be returned. By refreshing the workspace, the Transaction Reporter will continue the trace from where it last reached. In this way, a complete trace can be reached after providing some feedback to the workspace.
The following parameters are of interest:
- WORKSPACE_TRACE_TIME_LIMIT_SECONDS=0
- WORKSPACE_TRACE_DEPTH_LIMIT=0
Other similar parameters for Situation initiated traces are:
- SITUATION_TRACE_TIME_LIMIT_SECONDS=0
- SITUATION_TRACE_DEPTH_LIMIT=0
The following parameter improves Transaction Reporter performance when there are multiple Transaction Collectors:
- THREAD_POOL_SIZE=3
Accessing this workspace
Access this workspace by clicking on a category in the Navigator view, or by right-clicking a specific application, component, server, or transaction.
Links to other workspaces
When you have identified a problem, you can then link to the Applications Summary workspace or the Component Detail workspace by right-clicking an icon in the topology and selecting Link To.... You can also right-click a link icon in the table. This can help you determine where a problem started.