5 Tips For Websphere Application Server Monitoring

By Author: Joseph Mathew
Total Articles: 11
Comment this article

Many enterprises still rely on IBM Websphere servers to run mission-critical applications. However, websphere applications can be subject to performance issues that result in frantic finger pointing and long, tension-filled meetings—or worse, they an experience Tier-1 problems that results in serious performance and availability issues. Here are 5 tips for JVM management within a Websphere environment.

1. Ensure that you have 100% visbility

You can’t engage in productive web performance monitoring if you’re staring at a “black box” all the time. Make sure you have the right tools to truly see your environment—preferably a tool that discovers and monitors application code running on WebSphere and provides deep visibility of how it executes through the JVM in production. This enables users to understand where latency is spent and how system resource like CPU and memory is consumed by an application and business transactions.

2. Detect OutofMemoryExceptions and Memory Leaks

Engaging ...
... in websphere application server monitoring also means keeping on top of heap usage, garbage collection, and utilization of key memory pools over-time. It’s also necessary to track and monitor memory leaks automatically inside WebSphere Server and its JVM with minimal user analysis, allowing root cause to be found in minutes instead of days or weeks.

3. Monitor key server metrics

Best-in-class JVM management includes monitoring key metrics such as:

• Average Response Time
• CPU Utilization
• Disk I/O
• Thread Pools
• Enterprise Java Bean (EJB) Pool Statistics
• Java Database Connectivity (JDBC) Connection Pool Statistics
• Session Statistics
• Servlet Session Counts
• Custom MBeans attributes via JMX

4. Don’t forget the database

Web performance monitoring often focuses on the browser layer as well as the middlware websphere application server stack, but SQL can also be an important part of the mix as well. Effective websphere monitoring also detect both SQL statement and NoSQL query latency with full visibility into SQL text and NOSQL queries like Cassandra Thrift and MongoDB BSON. Support for all relational databases like Oracle, SQL Server, MySQL, Sybase and DB2 as well as NoSQL database like Cassandra, MongoDB, CouchDB and HaDoop.

5. Get to code-level fast – within minutes

When you do detect performance problems, you should be able to invoke a full call stack trace and get to code-level detail within seconds or minutes. Don’t rely on log files—that will take hours! Instead, you should engage in business transaction monitoring that enables you to “find the smoking gun” in the web performance monitoring issue in seconds.

The right application performance monitoring tool can assist you with all of these Websphere application server monitoring issues. Do your due diligence and find an APM tool that can give you all of these benefits, while running production at very low overhead.