5 Tips For Apache Monitoring

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

Apache monitoring environments can be fraught with hidden and frightening performance problems—everything from code deadlocks to failed business transactions to slow SQL calls. If you depend on log files to resolve these issues, you may find yourself spending far too much time in War Rooms and answering angry phone calls from both your co-workers and your end users. Instead of wasting time on Mean-Time-to-Resolution, engage in a true apache monitoring strategy that can resolve these performance issues quickly. Here’s 5 tips for doing so.

1. Map and Visualize Your JVM Dependencies

When performing server monitoring you should be able to visualize and map your Java virtual machines automatically discovering and mapping of all tiers that service and interact with your java Application and JVM. This gives you a far superior high-level view into your application performance than other Java performance tools.

2. Understand Your Key Metrics

You ...
... should also able to Monitor JVM health and Performance, understanding the health of your JVM with key Java profiling metrics like response time, throughput, exception rate, and garbage collection time as well as key system resource like CPU, memory, and disk I/O. You can should also be able to eliminate code deadlock Java issues.

3. Engage in Transaction Tracing

Think of business transactions as a “user generated request,” such as “Add to Cart” or “Check Out my eCommerce Order.” Tracing business transactions is the absolute best way to ensure world-class apache monitoring, because you’re monitoring what really matters to the end user. In addition, the notin of the business transaction is easily understood by everyone in the business, including the CIO!

Distributed Business Transaction Tracing enables an advanced tag-and-follow tracing of distributed transactions across your java virtual machines. You gain a simple, intuitive view of how a distributed business transaction executes and where bottlenecks exist. And finally, you can troubleshoot Java Code Latency in Minutes, getting full visibility of Java code execution as business transactions and requests are processed by the JVM. This allows you to isolate latency and pinpoint its root cause in minutes.

4. Keep Up with JMX Trending

In regards to JMX trending, you can get a comprehensive metrics viewer that allows you to correlate, trend and analyze key metrics from your JVM and Container (via JMX) as well as metrics via custom MBeans, as well as detect and troubleshoot memory leaks and object thrash issues, even under production load.

5. Don’t Neglect SQL Monitoring

It’s an unfortunate truth that often slow SQL calls are the culprit in regards to server monitoring issues. The right Apache monitoring tool can provide you with monitoring capabilities for both SQL (Oracle, SQL Server, MySQL, Sybase, DB2, …) and NoSQL technologies (Cassandra, MongoDB, and Haoop).