Originally published here by Todd Kindsfather, February 1, 2017.
Digital applications are the lifeblood of business today. They are the primary means of interaction with customers. Itâ€™s imperative applications are always available, provide optimal performance and deliver exceptional customer experiences. If not, you can probably expect your customers to say goodbye.
Applications run on top of a complex web of software components. Some provide the platform and connectivity needed to deliver services and others move data between devices. IBM provides a robust set of components, including WebSphere Application Server, MQ, IBM Integration Bus, Datapower and more. Each helps deliver numerous applications.
To ensure your applications perform optimally, you should manage the health and welfare of software components. IBM Application Performance Management (APM) monitors the performance and availability of your critical IBM applications to identify problems before impacting users, visualize performance bottlenecks and more.
Fine-tune your software with APM metrics
Many people think of monitoring as being alerted about a problem and guided to the issue source to fix it. But another motivation for monitoring is to proactively avoid those problems in the first place. Adopting DevOps methodology, you can take information from monitoring to your developers for fine-tuning to improve application performance. You can gather metrics at short intervals, making it more likely that youâ€™ll spot trends or anomalies that indicate bottlenecks.
APM allows you to monitor the key metrics of your IBM software environment for optimal behavior. But itâ€™s also important to measure the CPU, memory and network utilization to ensure bottlenecks arenâ€™t at the platform level.
To illustrate, here are some of the metrics you can use from APM to tune WebSphere Application Server:
â€˘Heap utilization and garbage collection statistics to determine if memory leaks are occurring
â€˘Database Connection pools to identify if they are too small to handle the load that is placed on them
â€˘Thread pools to determine if they are too small to handle the load
â€˘Web Services â€“ identify most used web services and performance problems, including if it is a code or underlying resource problem
Speed resolution with APM metrics
You can also use APM to identify problems and speed up resolution time. It works similarly to fine-tuning, but with more frequent metric gathering and alerting rather than reporting. This is also where IBM APM outshines other solutions, offering quicker troubleshooting and resolution.
Transaction tracking can dramatically improve problem diagnosis by isolating the source. This ensures the issue is routed to the right SME and does not involve others responsible for different areas of the application environment. IBM Operations Analytics â€“ Predictive Insights automatically determines baselines for metrics and will alert you about deviations from that baseline. It can also identify related metrics helpful for quicker identification of problemsâ€™ cause.
If youâ€™re using IBM components to build applications, you should consider coupling them with IBM APMâ€™s monitoring designed to help tune those components for optimal performance and quick problem resolution. Result? An optimal customer experience.
Want hands-on experience with IBM APM? Attend IBM InterConnect for countless sessions, labs, and educational opportunities.