Send all X-Ray

1
Send all X-Rays to appman_at_appdynamics.com All
X-Rays received by App Man will be scrubbed of
any Customer Names or Identity using his super
powers so all content and screenshots are
anonymous. Your X-Ray(s) may be used by App Man
in future as content for his weekly blog. You
may submit more than one submission per
bottleneck or problem. 125 Amazon gift cards
will be sent immediately via email on receiving
an X-Ray. Winners of the 50 TV worth 1500 will
be announced every 6 months. Good Luck and may
Appmans X-Ray Byte Code be with you!
2
Example x-ray report By App man
ltEdit the following slides to document the
bottleneck or issue you solved using AppDynamics
Progt
Appman files
3
Application Summary
Purpose of Application Industry Architecture
Type Technologies Instance Count Avg Daily
User Count Peak Transaction Throughput Avg
Daily Revenue of Unique Business Transactions
E-Commerce Online Bookstore Retail SOA Tomcat
(3), MQ (1), MySQL (4), 60,000 users 1,250
transactions per minute 750,000 63
Appman files
4
Problem Summary
Middle of May 2011 AppDynamics Alert Our
checkout business transactions were taking up to
10 seconds to complete. Inventory Web Service
was bottleneck due to slow SQL statement 1
hour 15 minutes I wouldnt like to guess, few
hours maybe?
When did the Problem Occur? How did you learn of
the Problem? Brief Description of
Problem Brief Description of Root
Cause Time to Install X-Ray Vision
Resolution Time with X-Ray Vision? (Est)
Resolution Time without X-Ray vision?
Appman files
5
RESULTS Summary
We were able to process Checkout business
transactions faster with less risk to our
customers. A slow user experience might cause our
customers to abandon the final step. We also
noticed our Database resource usage dropped after
the DBA tuned the SQL statement. Powerful in
that it lets you go from high level to low level
in a few clicks to find the root cause.
What impact did solving this problem have on your
business? How would you describe App
Mans X-Ray vision?
Appman files
6
Screenshot 1
I logged into AppDynamics and noticed several
problems in the Application Dashboard view.
Appman files
7
Screenshot 2
The Problems screen showed me the Checkout
Business Transaction has been running slow so I
drilled down.
Appman files
8
Screenshot 3
AppDynamics captured 3 snapshots of the Checkout
business transaction which highlighted a 10
second response time. I then drilled down in to
one of these to troubleshoot.
Appman files
9
Screenshot 4
Here is the transaction flow for the 10 second
Checkout business transaction. I then drilled
down on the start tier.
Appman files
10
Screenshot 5
I clicked on this link to drill in context to
understand where the 10 seconds was spent.
Here is the complete call stack of how the
Checkout transaction executed in the E-Commerce
Java tier. Notice how all the time is spent in a
web services call.
Appman files
11
Screenshot 6
Here we see the full call stack of how that web
service executed in the invoked Inventory JVM.
Turns out the 10 seconds was caused by a slow
running SQL statement as shown.
Appman files