Balrog/Meetings/CloudOps - January 18, 2017
From MozillaWiki
Attendees
bhearsum, mostlygeek, chartjes, ckolos
Notes
- Running new extraction script on prod db, or clone?
- mostlygeek will run on production with read only credentials when ready
- Contract tests
- chartjes working on them
- Will be set-up on QA jenkins and kicked off as part of deployment
- Athena/Log parsing
- We have 1.6TB logs/day
- Not efficient to use Athena - ~$5/TB/query
- Someone from ops is looking at converting raw logs into Apache Parquet format
- Probably will get ~99% improvement in efficiency
- No loss of info or granularity
- Should be able to enable Athena afterwards