The last app monitoring tool review on our tour is AppDynamics. There’s more than meets the eye with AppDynamics. In terms of features, AppDynamics is huge in both depth and breadth. The app has potential for use in large firms who want to keep track of everything going on in and with their application. This is because each account can have multiple users monitoring multiple web apps.
Setup
I noticed something strange when I began to install the agent. I didn’t have the option to select the PHP agent! This was unexpected because AppDynamics says that they support PHP applications (Magento in our case). I had to get the PHP agent package from another AppDynamics account in order to get the agent working. The issue caused us to be late for the demonstration web meeting. Monitoring agents should be easy to install, but their support team was not as helpful.
Application
Once the agent was installed, we were on our way to exploring AppDynamics. The first page shows all the different applications being monitored. In our case MAGE_TEST is the group of servers housing the applications we’re testing.
After delving deeper into the MAGE_TEST application, we are greeted with the application dashboard. In the dashboard, I uncovered the Application Flow Map. While it may be superficial, it is one of my favorite features. It lays out your application network’s topology and basic metrics in a way that is easy to understand.
The Application Flow Map for our environment (click to expand)
Drill Down into a Web Transaction
Past all the eye candy, there is another tab that peaks our interests. Clicking on Top Business Transactions neatly sorts the transactions it sees as most important. They include:
- Transactions by Load
- Transactions by Response Time
- Transactions by Errors
- Transactions by Slow Transactions
- Transactions by Stalls
- Transactions by Health Rule Violations
We notice business transaction catalog : product : view is taking 17.9 seconds. When we double click it, we are shown the Application Flow Map associated with just that transaction. In this case, it’s just NODE_3 and the shared RDS database. From here, we select a transaction snapshot to “Drill Down” into. These show the slowest pages related to our transaction. Finally we get to the call graph which is similar to NewRelic’s transaction trace. The call graph shows us the slow code in the app. This can help developers pinpoint performance bottlenecks in the application layer. The Hot Spots section shows the slowest methods of that transaction. In addition the SQL Calls tab shows queries sent to the database. This feature didn’t really exist in NewRelic or AppFirst, so it was nice to see for the first time in AppDynamics.
Other Features and Impressions
We’ve only gone over a few of the many features in AppDynamics. There are other features we liked but aren’t elaborating on. They include the following:
- The ability to save time windows
- The all encompassing Metric Browser
- Scalability Analysis reports
- Following transactions through the network topology (Inception of app monitoring)
- SMS notification alerts
In general, AppDynamics is still a thorough and impressive app monitoring service. Despite the rocky start, the application remains a useful tool to anyone who wishes to optimize the performance of their web application.