Tuesday, June 6, 2017

ODATA metrics, statistics & analysis

Intro

Usually SAPUI5 project life cycle includes different phases, starting from specifications and going through ODATA and UI development.

As projects evolve, question of usage and analytics will arise:
  • How many users do we have in our SAPUI5 application? 
  • How often do they log on and use the solution? 
  • What are the ODATA entities used during application usage? 
  • How often do users perform create & update ODATA operations?
  • Which entities were created through the SAPUI5 application?
In this short blog I will try to cover general SAP NetWeaver Gateway tools that allow us analyzing the usage & efficiency of our custom SAPUI5 developments & standard Fiori apps.

Detailed metering data

By executing report /IWFND/R_METERING_VIEW through SE38 you will be able to get all ODATA calls information by user & date range.





You can use embedded filters & sorting to try and analyze the data, in case you need to get your hands on the actual data - you can access it by going to table /IWFND/L_MET_DAT

Aggregated data


Table /IWFND/D_MET_AGR will show you 2 important metrics:

  • Usage per CRUD call (create / read / update / delete)
  • Number of active users per month
 

Table /IWFND/L_METAGR will show you aggregated data on service / month level



Application log

You can also use the application log to see errors & execution times in case you wish to analyze errors / loads of your ODATA services by launching SAP Gateway Application Log Viewer through  transaction /n/IWFND/APPS_LOG

Error log

All failed ODATA requests are written into error log, which is available through transaction /IWFND/ERROR_LOG 

Jobs behind the scene

Once you activate the SAP NetWeaver Gateway you are automatically scheduling 2 standard jobs:
  • SAP_IWFND_METERING_AGG
    This job is scheduled daily and updates the aggregation summaries
  • SAP_IWFND_METERING_DEL
    This job is also scheduled daily, it cleans up aggregation data older than 2.5 years back, and also cleans up the data from  /IWFND/L_MET_DAT on monthly level.

    In case you execute the report /IWFND/R_METERING_VIEW by 1st of the month you are going to see 0 entries, so it might be a good idea to cancel this job scheduling to collect your data.