APM- Concept (POC) to Implementation

Santander’s existing solution to measure performance matrices would not go far enough to monitor and manage application performance. Data collected from various generated logs were not aggregated and it was not easy to create meaningful dashboard for various stakeholders. It’s way too cumbersome (if not impossible) to correlate the activities logged in different tiers in multi-tiers applications. Also, we didn’t have the tool to alert responsible parties when certain performance thresholds are exceeded.

APM solution is a must have to effectively monitor and manage application performance. APM solution would allow an enterprise quickly detect and diagnose performance problems once the defined threshold is crossed.

The Bottom Line: Know the problem as they occur, identify them in near real time and take proactive actions before the end users ask for help. APM solution will allow you to achieve those capabilities.

So, what is APM? It stands for Application Performance Monitoring!!

• Identify and solve the production and non-production issues as they occur or before too many users are impacted.
• Provide end to end visibility of events/transactions within applications and correlate them for effective diagnosis.
• Dashboard to show the heath of the enterprise which can be drilldown into details by process or application.
• Control access to dashboard with Windows Integrated Authentication and AD/LDAP groups for Authorization.
• Allow masking of sensitive data collected by APM agents.
• Alert if response time of requests exceed the defined threshold (N seconds/minutes).
• Temporary exclusion of alarm/alert. This is to allow exclusion to known exceptions or to eliminate false positives.
• Agents collecting and feeding the data to APM management server must not consume more than 2% of the host CPU resource.
• No need to make any change to any applications in any tier.

Every organization is different and criteria would vary based on the existing architecture in the organization. There is no one size that will fit all! Some company may choose to go vertical (scale up) to add more computing power as opposed to horizontal (scale out). You can’t change the fact but that can impact licensing cost when license is counted per machine or CPU. That’s just one example. There are certain criteria those would carry more weight because those are must have capabilities. The rest are good to have and we would give them less weight. Again, I can’t disclose the competing company names because of non-disclosure agreement. They all were good with slight variation based on the stated criteria. Of course, one came at the top based on its weighted average merits. My role was to show the facts to IT management and it’s up to them to make the final decision.

Of course, there were alternative to APM product but how effective were they in problem resolution? It’s all about bringing new capabilities to the organization and to empower the people who need them most- primarily for IT Support groups who can effectively troubleshoot operational issues in prod and non-prod and for Application Developers who can identify things that are not caught/logged at the applications. I thank Santander’s IT leaders for trusting me to lead the APM initiative from evaluation to implementation and, finally, transitioning to support group for it’s up keeping and maturation. I have had the opportunity to evaluate competing solutions based on prevailing Information Systems Architecture and Technology Architecture (talking about hardware, OS, software stack (.NET primarily)) and I can’t name them due to nondisclosure agreement. I led the implementation effort with a defined scope. You can’t bring all apps under APM in one iteration for various reasons including resource limitation, minimize risk/impact and to achieve a learning curve. It’s been over a year now and multiple stakeholders are using DynaTrace APM today.

APM technology is science and it can do its job but bringing all the stakeholders together from their comfort zone is whole new territory- it is an art! I must thank my colleagues and IT leaders who supported me in the adoption of APM. Implementation of APM is yet another success story at Santander and I am happy to be part of it.