It is the fifth straight year that I have been writing Tech Tips. I have learned so much on the journey from interacting with many of you and want to convey my thanks.
This is a special tech tip since it is number fifty. [It could be number forty-nine or fifty-one. But close enough :-) ]
In previous tips, the Transaction Definition Lifecycle was covered in detail. But one area was left out and that is the planning phase. This is covered in part in a new KB:
But this KB does not include an overall framework. A modest start of doing this is provided below.
First Step: What and Why am I Monitoring?
One or more applications are to be monitored as part of a company's application monitoring strategy. Applications may be prioritized based on relative health, political concerns, application complexity, and business/financial focus. It is important to articulate why you are monitoring and make sure your implementation plan follows that.
This can be due to reasons similar to the above selection criteria:
- Desire for high customer satisfaction.
- Lost revenue concerns.
- Improve/increase application availability
- Eliminate potential problems for new releases
The above reasons will determine:
- The business transactions/transactions to be monitored.
- The metrics to be monitored
- The reports to be generated.
Second Step: Planning the Business Transaction Recording Lifecycle
Once the what and why is completed, then the traditional project management planning stage is initiated. This includes
- What are the roles and tasks for the various lifecycle processes?
- Who will monitor the monitoring for overall success?
- Will reporting used reactively and/or for capacity planning? For defect forecasting?
- How integrated with APM CE Monitoring be with the application development lifecycle?
Take the time to do this step right including capturing assumptions, risks, using dedicated resources, have clear tasks, and realistic timeframes.
Step Three: Know your Application and Diving In.
The better you understand the application, the less likely that you have issues. This includes the following:
- Complexity of application
- Simple -- HTTP/HTTPS Request-based traffic, ASP/ASPX
- Intermediate - Complex -- XML/Web Services, FLEX, JSON, SAP, Seibel, Microsoft Web Applications, Oracle Forms,
- Infrastructure setup (Web servers, load balancers, firewalls, reverse proxies, network switches)
- Location of TIM in relation to the application's server.
- Application's use of SSL (cipher suite TLS version), and other protocols.
- Compliance to HTTP Protocol(includes content-type, accept language).
- Verify round trip affinity (The transaction request and response goes through the same set of servers.)
- Sitting through walk-throughs with the developers on key transactions including determining how to identify users/sessions, group users, key parameter-value pairs to identify
transactions, possible overlapping transactions.
- Go through several dry runs on recordings/monitoring to see if what development said above will work or if another HTTP field needs to be added to the application.
Thank you all for your support and feedback. The quest for promoting/achieving scalable and stable APM infrastructures goes on. I plan to keep monthly tips for as long as I can.
Questions for Discussion:
1. Are you doing the above planning process? If so, please share the details.
2. Which one of the fifty tech tips was your favorite?
3. What other topics would you like me to cover?