Failover adding 250ms to simple query

[Blake Acheson]

I have a very simple query that updates a field using a primary key. When running a single instance the query completes in 50ms. When adding failover the same query completes in 300ms. This basically cripples our performance. Is this standard behaviour?

[Carlos (Cloud Platform Support)]

I believe you could be affected by replication lag as explained and addressed in this document. Semi-synchronous replication might have an impact. Using SSD disks and making sure that your instances have enough SSD and RAM available might help. I found a related issue you might want to have a look at.

[Blake Acheson]

Thanks Carlos. I tried to create a new failover and it's in an "unavailable" state. This happens almost every time I interact with a failover these days. Contacting support.