refresh_queries shouldn't break because of a single query having a bad schedule object#4163

Merged

rauchy merged 9 commits intomasterfrom resilient-refresh-queries

Mar 1, 2020

+216-273

Commits on Feb 25, 2020

move filtering of invalid schedules to the query
Omer Lachish
committed
simplify retrieved_at assignment and wrap in a try/except block to avoid one query blowing up the rest
Omer Lachish
committed
refactor refresh_queries to use simpler functions with a single responsibility and add try/except blocks to avoid one query blowing up the rest
Omer Lachish
committed
avoid blowing up when job locks point to expired Job objects. Enqueue them again instead
Omer Lachish
committed
there's no need to check for the existence of interval - all schedules have intervals
Omer Lachish
committed
disable faulty schedules
Omer Lachish
committed

Commits on Feb 26, 2020

reduce FP style in refresh_queries
Omer Lachish
committed
report refresh_queries errors to Sentry (if it is configured)
Omer Lachish
committed

Commits on Feb 27, 2020

avoid using exists+fetch and use exceptions instead
Omer Lachish
committed