Hosted on heroku with a hosted heroku postgres dev instance, I observed a drop in the db response time : - from a consistent average of 10ms over the last week

Hosted on heroku with a hosted heroku postgres dev instance, I observed a drop in the db response time : - from a consistent average of 10ms over the last week - to a new consistent average of ~2.5ms beetwen 17:31 and 17:35 UTC. AWS started to report the current issue at 17:38 UTC. My app then experienced some intermittent issues (reported by newrelic pinging it every 30s). Don't know if it's related, could it be some sort of hw upgrade that went wrong ? I did a push affecting my most used queries but that was one and a half hour sooner, at 15:56 UTC, so probably unrelated.