AppDynamics Application Intelligence Platform

3.8.x Documentation

PDFs

Videos

Release Notes

 

How You Know When Response Times are Slow

There are many ways you can learn that your application's response time is slow:

  • You received an email or SMS alert from AppDynamics (see Alert and Respond). The alert provides details about the problem that triggered the alert.  
  • Someone reported a problem such as "it's taking a long time to check out" or "the app timed out when I tried to add an item to the cart."
  • A custom dashboard shows a problem. 
  • An AppDynamics dashboard shows a problem. In AppDynamics, look at:
    • Business transaction health metrics in the Transaction Scorecard pane.  The bar charts show slow or stalled transactions as detected based on the configured thresholds.

    • The Response Time graph. If you see spikes in the graph, the problem is slow response time.  

    • Events in the Events List. Scroll the list and look for red or yellow icons related to performance health rules based on response time. 
    • Traffic flow lines in a flow map. If you see yellow or red flow lines the problem is likely to be slow response time.

    • Business Transaction Health status in the Business Transaction Health pane. If you see yellow or red bars, there are health rule violations that may affect response time.

    • Tier icons in the flow map. If the icon is yellow or red there is a problem with one or more nodes that may affect response time. 
    • Health rule status in the Server Health pane. If you see red or yellow, the problem reflects a server health rule violation that may affect response time.  

Basic Troubleshooting Techniques

At any time you can click Troubleshoot -> Slow Response Times and the Slow Response Times window opens showing two tabs. You can drill down into transaction issues in the Slow Transactions tab, and into database or remote services issues in the Slowest DB & Remote Services tab.

Slow and Stalled Transactions

There are many reasons why a business transaction may be slow or stalled. The Slow Response Times tab helps you find the root cause whether that be resource or thread contention, deadlock, race condition, or something else.

By default AppDynamics considers a slow transaction one that lasts longer than 3 times the standard deviation for the last two hours and a very slow transaction 4 times the baseline for the last two hours.

By default AppDynamics considers a transaction that lasts longer than 45 seconds (4500 milliseconds) to be stalled.

You can configure these thresholds to better match your environment. See Thresholds and Configure Thresholds.

This interactive video traces the typical steps of identifying the cause of a slow transaction.


To troubleshoot slow and stalled transactions

1. Click Troubleshoot -> Slow Response Times.
You can also access this information from tabs in the various dashboards.

2. Click the Slow Transactions tab if it is not selected.

  • In the upper pane AppDynamics displays a graph of the slow, very slow, and stalled transactions for the time period specified in the Time Range drop-down menu. If the load is not displayed, you can click the Plot Load checkbox at the upper right to see the load.
  • In the lower pane AppDynamics displays the transaction snapshots for slow, very slow, and stalled transactions.

3. In the lower pane, click the Exe Time column to sort the transactions from slowest to fastest.

To drill down to the root cause of the slow or stalled transaction, select a snapshot from the list and click View Transaction Snapshot. See Transaction Snapshots.

Slow Database and Remote Service Calls

Although AppDynamics does not instrument database and remote service servers directly, it collects metrics about calls to these backends from the instrumented app servers. This allows you to drill down to the root cause of slow database and remote service calls.

To troubleshoot slow database and remote service calls

1. Click Troubleshoot -> Slow Response Times.
You can also access this information from tabs in the various dashboards.

2. Click the Slowest DB & Remote Service Calls tab if it is not selected.

  • In the Call Type panel, you can select the type of call for which you want to see information, or select All Calls. 
  • The Call panel displays the average time per call, number of calls, and maximum execution time (Max Time) for the calls with the longest execution time.
  • In the lower panel, you can see details or correlated snapshots for the selected call.

3. Sort by Average Time per Call to display the slowest calls at the top of the list.

4. To see transaction snapshots for the business transaction that is correlated with a slow call, you can:

  • Click the View Snapshots link in the right column to display correlated snapshots in a new window.
  • Select the call and click the Correlated Snapshots tab in the lower panel to display correlated snapshots at the bottom of the screen.

5. Click the Exe Time column to sort the transactions from slowest to fastest.

To drill down to the root cause of the slow call, select a snapshot from the list and click View Transaction Snapshot. See Transaction Snapshots.

Learn More