· Find out the solid requirements/needs for a strong APM (whether we already have it or not)? And WHY?
· Find out the work has already been done, what frameworks exist today and what needs to do for fulfilling the requirements
· Review the APM and the frameworks of Application Performance Management
· Review the problem we are facing today as we don’t have the APM
· Review the solution of the current problem and Architecture thought on this
· We will not be discussing the tools today Discussing requirements.
· Review the workes already been done and review the framework if exist
· Come up with the list we need to metrified
· Any remaining questions?
Recap of APM
APM or Application Performance Management is the monitoring and management systems of the availability and performance of the applications.
According to the APM Conceptual Framework, Performance Management has five phases
End User Experience: Its referred to as real-time monitoring and goal of the end-user experience is to make sure the communication between a client/user request and server response is efficient and effective.
Runtime Application Architecture: It is a Graphical view of application components or different application who works together to deliver business functionality.
Business Transactions: This functionality track the user request within the application across all of the components that service those request.
Component deep-dive monitoring: The functionality of this phase is to record and measure the internal workings of the application components.
Analytics: Analytic should help to identify and solve the problem based on the metrics we can collect from the application
Problem We are facing today ?
· Using Splunk for everything
· Relying on Splunk alert for knowing if something break
· Not having the many options to triage the issue and find the root cause
· Not able to trace end to end customer request
· No visuality on different component dependency
· Not using the Operational Metrics
1. CPU Uses
2. Memory Uses
· Not Using the Application metrics
Work has already been done
· Using the InfluxDB database to collect the matrics
· Added Metrics framework in the