We've observed that EL database (m4-master.equiad.wmnet) has some "holes", some time spans that contain no data for any event table. For instance:
2015-04-09 17:20 -> 18:35 2015-04-11 03:35 -> 05:20 2015-04-11 14:00 -> 15:20 2015-04-11 19:00 -> 20:00 2015-04-12 14:30 -> 15:40 2015-04-13 11:35 -> 12:30 2015-04-13 16:30 -> 17:35
We know that:
- EL processor is validating correctly the missing events and writing the correspondent logs.
- EL mysql consumer is homogeneously writing events at the time of the holes. But we can not tell if these events are the ones that are missing or not, because the consumer logs indicate only the insertion timestamp and not the event timestamp.
- The database does not receive any inserts for the missing data.