This past Monday, Google Cloud SQL experienced an issue that caused our database and many others to go down for 3 hours. You can read more about this here: https://status.cloud.google.com/incident/cloud-sql/18001
In June, we had migrated to Google Cloud SQL to reduce operational overhead so we could reallocate that time to product development. However, in the years our product has been live, we have never experienced this much database downtime when running our own database server.
This is unacceptable to us. As such, we will be migrating our database back to our own server (with 20x the power!) Saturday at 8:30pm PT. We believe that less than an hour of downtime now is better than an unpredictable amount of downtime at unknown points in the future.
Due to many hours spent on improving our database over the past month, this migration should take only a fraction of the time the previous migration did. We are scheduling one hour but expect it to take less than that. It is important to us to perform this migration before the university semester schedule commences next week which is the cause of the short notice.
Thank you for understanding as we work hard to ensure you and your users are delivered high quality, stable products.