1

I am writing a functional design for enhancing our existing application logging. Our Core application is built on C and we have a integration layer on top of that to talk with other third party systems. Until now, we do not have a centralized logging mechanism where one would be able to track the entire transaction with the help of one Unique ID. After implementing this design, we are hoping we would be able to achieve it.

We have created 5 different database tables:

  1. Audit_Log
  2. System_Event_Log
  3. Debug_Log
  4. Err_Log
  5. Batch_Log

As these tables gets populated we will hand over the SQL query to Foglight for application performance monitoring.

Question: How do I write use cases for this implementation before I hand it over to my QA. Because I have to include all the integration points in my FD.

I would welcome your thoughts on it.

Improve this question
2
  • Do you know what your requirements are?
    – user246
    Commented May 27, 2014 at 18:51
  • The requirements are straightforward, we should integrate with foglight which would probably run the SQL queries we have and access our application's performance. For example, our processing time for each transaction and whether we meet the SLA or not.
    – alphadom03
    Commented May 27, 2014 at 19:09

1 Answer 1

Reset to default
0

My perspective of this question is in terms of scenarios and how it will be handled. There are different use cases which may fall in this criteria. Logging on different tables for different scenarios

  • Example #1 - Disk Space issue, When there is insufficient disk space available for MSSQL, Other processing this would go into Err_Log, System_Event_Log
  • Example #2 - When there is timeout of operations (network, I/O issues it need to be captured)
  • Example #3 - When the CPU exceeds 80% it need to be captured and logged in System_event_log
  • Example #4 - When there is error in processing (Invalid data formats, error thrown by application) it will be logged in error_log
  • Example #5 - Archival mechanism of this logs (How frequently they needed to be cleaned up)
  • Example #6 - If the same error is occured multiple times do you increase severity / prompt for immediate attention
  • Example #7 - Identity failures, frequency of failures for Errors
Improve this answer
2
  • Thanks Siva. These log tables are for application logging purposes. We understand the non functional requirements, and that has been taken care off. I am trying to write use cases to test the logging framework we have designed. While logs can be used to troubleshoot or debug while testing, to write use cases to test a logging framework looks like a challenge.
    – alphadom03
    Commented May 28, 2014 at 11:38
  • There are two aspects. Testing how you have implemented logging - validate implementation, secondly use cases to validate logging. Implementation testing depends on your design (how frequently you log, log formats, events associated with it).
    – Siva
    Commented May 28, 2014 at 13:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.