CA Tuesday Tip: APM CE Urban Legends and Misconceptions
Throughout the years, there has been various misconceptions and tall tales told about APM CE (CEM). But no one has attempted to bring these to the light of day until now.
Legend: TIM and TESS are names of some developer's child, wife, or other relation
Status: Unconfirmed. TIM and TESS have always stood for various acronyms such as Transaction Impact Monitor and Transaction Events and Statistics Server. But there is a
rumor that TIM was originally called probe.
Legend: CEM once had a Basic Recorder in addition to the current Advanced Recorder
Status: Confirmed. In early releases, CEM had a recorder that could capture transactions from the browser. It was called the Basic Recorder.
Legend: You once created a 600+ page book on CEM.
Status: Confirmed. Work was halted due to extenuating circumstances. A small part of it has ended up as Tuesday Tips.
Legend: Timestock is the original name of a science fiction novel about business and time travel.
Status: Unconfirmed. It may be a book. But it was also the name of the original company that created CEM.
Misconception: Success Rate is for Slow Time only. Missing and Partial Responses incorrectly bring down success rate.
Success Rate is for ALL defect types. That includes missing and partial transactions, unauthorized logins, client/server errors, etc. Some of these may be false positives due to bad network data. So if you do not like the success rate numbers that you are receiving, try to figure out why the defect is being created. A temporary measure may use Yield as a metric instead.
Misconception: E-commerce User Groups can be set up on the same type of user group identifiers as Enterprise Mode.
E-Commerce User Groups in Reports are typically only for Unspecified Users. But using HTTP request attributes, other user groups can be set up. However, manual IP subnet e-commerce user groups are not supported for statistics and reports.
Misconception: Duplicate Session ID is an error message.
It is a warning message and typically a concern because it fills up logs and the APM CE Events screen.
Misconception: There is no wrong way to set up a TIM to see network traffic. I shouldn't worry about dirty traffic.
There are many things that can be set up incorrectly resulting in not seeing traffic or seeing misleading traffic. These include
- Span, Tap, or Aggregator is misconfigured so it is seeing router/non-HTTP or one-way traffic. They also can be filtering traffic from desired web server.
- The ifconfig files on the TIM are misconfigured
- No network interface is enabled for TIM monitoring
- Private keys are incorrect or have wrong passphrase
- Dirty network traffic can either result in not seeing transactions, false positive defects, SSL decode failures, and many other factors.It should be eliminated if possible.
Misconception: Thinking that CEM metrics and Introscope CEM-based metrics measure things the same way.
There are a variety of factors that may impact differences in metrics. This include: reporting on different timeframes (such as CEM can only show hourly data and Introscope can show data in seconds and minutes), CEM being concerned about web servers and Introscope about application servers and backends, and more. This issue is covered in the knowledgebase and other documents.
Questions for Discussion:
1) Are there other APM urban legends and urban legends that you wish to share?
2) Do you want another Tip on this topic?
3) What other troubleshooting topics would you like to be covered in Tuesday Tips?