What happened

One of our DB instances had a downtime for about 3 minutes. This affected our API capabilities to respond properly to requests.

What went wrong

A code change introduced a feature that wasn’t cached hence heavily increasing requests to the DB.

What we are doing about this

• Upgrade database to reduce the downtime (in case it should ever happen again, from 140 seconds to 60 at most).
• Configure proper maintenance schedule so we have better control of automated maintenance by cloud provider
• Improve caching so even if database is down we can still serve the traffic