We want to apologize to the Feng Sky customers who suffered a four hour downtime yesterday.
One of our main Database servers has been presenting performance issues for the last three weeks. We were slow at better addressing the problem earlier. We attempted several solutions but none was sufficient. The issue got to a point yesterday when we needed to move many customer Databases to a new server.
The issue has been taking the team many straight hours of work since yesterday.
All customers should be up and running now. Please let us know if you are experiencing any problems.
A few trial accounts are still experiencing troubles. Trial accounts follow a different path through our servers, so fixing those accounts is taking us a bit longer.
We have prioritized fixing the issue over producing a more detailed report of the actions we are following and the problems that caused this downtime. After we ensure that all services are back to normal we will post a more detailed update on what happened.
We are very sorry for the inconvenience. We promise we will learn from this bad experience and ensure even greater measures to prevent this kind of issues.
Thank you for your support and understanding.
—
Conrado Vina
CEO
Feng Sky October 05 downtime
Consider implementing PostgreSQL as a backend to FengOffice. MySQL got the momentum it has for historical reasons, but when concurrency increases it behaves embarrassingly bad, whereas Postgres scales gracefully.
Thanks for the tip. Definitely something to consider.
We are still assessing the problem. I believe we can have an infrastructure architecture that should perform without any problems at even greater scale.