Last modified: 2014-02-26 12:53:42 UTC
We are starting to see potential instrumentation requests where: 1) the data collected might be unsuitable to be stored into SQL due to (a) the volume of events logged and/or (b) the fact that the analyst may not be using SQL at all to crunch this data. 2) the data collected needs to be segregated from other logs because it's subject to different privacy terms (for example the Donor Privacy Policy) or data retention conditions (per the Data Retention Guidelines). Should there be flags in JSON schemas deployed in production to automatically determine what subscriber listens to a specific log?
[moving from MediaWiki extensions to Analytics product - see bug 61946]