Wily Solution Scenario — Reactive Application Management
A large, worldwide telecommunications firm has chosen to implement
the Wily solution to better ensure the performance and
availability of its DSL reseller application. The firm chose to
implement Wily based on the capability of Wily’s management products and
on the solution’s ability to bridge the functional silos across
the IT organization. The application is expected to deliver
superior performance and conform to the firm’s high QoS standards,
which include continuous application improvement and adherence
to strict Service Level Agreements.
The firm’s Application
Support team uses Introscope®
Alerts, Dashboards and Blame™ technology to pinpoint problems
with the application itself, the application server and supporting
systems. Shortly after the application is deployed, the Application
Support specialist finds on his Introscope dashboard that
memory consumption has been steadily increasing. He suspects
that a memory
leak exists. He turns to Introscope LeakHunter™ and quickly
finds that the source of the memory leak is a specific EJB
that keeps
adding objects to a hash table without ever removing them.
The Application Development team fixes the problem and the
application is redeployed.
Using Introscope PowerPack™ for the Application
Server, the
application support specialist discovers that the JDBC connection
pool is running low, causing a slowdown in JDBC operation
response time. He informs the application server administrator
who increases
the JDBC connection pool, thus improving performance right
away.
With help from Introscope, the Application Support
team has successfully met the application’s SLA (90% of application
transactions must be completed within 6 seconds). But the
Office of the CIO recently redefined its QoS strategy and
has determined that the application must meet a new SLA that requires
95%
of the
transactions to be completed within 3 seconds.
The team uses Introscope Transaction
Tracer™ to track individual “rogue” transactions
that, while constituting a small percentage of overall transactions,
are responsible for causing the largest percentage of application
performance problems. Throughout the day, Transaction Tracer
identifies 152 rogue transactions. 140 of these transactions
utilize an EJB
that was supposed to have been removed. The application support
specialist notifies the development team who fixes the problem
and rolls out an update.
The remaining twelve rogue transactions
each take a different path through the application. The Application
Support specialist
sorts these transactions by response time. One by one, he
examines each transaction through Transaction Tracer’s patent-pending
SpeedTrap™ technology to find where the most time was spent
in the transaction path.
He opens the Transaction Tracer screen to view each individual transaction.
It shows a component-level breakdown of the transaction from
the application entry point through to the back-end requests.
A majority of the rogue transactions spend the most time making database
calls,
which are shown as the actual SQL statements that the transaction
makes. The Application Support specialist calls the DBA to
examine the SQL statement issues. The DBA then uses his own
Introscope
Workstation
to see his own views of the transactions, as well as other
application/database interaction metrics.
The application is
ready to be redeployed. By using Introscope and Wily’s family
of Java application management solutions to achieve high
application performance, the Application Support
team can now meet the new corporate QoS goals and ensure
consistently superior experiences for every user. |
Community
Partner Access
Site Map
Contact
Community
Partner Access
Site Map
Contact