Run Cluster on Spot instances?

So for last 2 weeks only one node had been restarted, this wasn’t a primary, so there was no effect on system. I also tried to restart primary manually, it didn’t cause any downtime, and in fact even the avg. response time, error rate and other vital metrics were not affected. We are able to serve more traffic faster and this positively affected our revenue.
So I believe this setup could work. Let’s give it another couple of weeks.

1 Like