Update: I got my data back! RDS is very confusing (what else is new with an AWS service?).
I’m building this cool app. It scrapes a bunch of interesting data, does some machine learning and natural language processing magic, blends in some user-generated content, and… does something useful. It’s a work in progress.
Over the last two months I’ve been stuffing my Amazon RDS Postgresql instance with a daily scrape of raw data, plus thousands of manually parsed pieces of training data. It’s ~~on the cloud~~, but don’t worry! I have daily snapshot backups!
I couldn’t connect to my database this morning. A quick call with support revealed that the credit card on file had been declined, and I was past due on charges. Twenty dollars worth! The support rep un-suspended my account for another month, I added a new card and payed off the charges, and was back in.
My RDS instance had vanished with barely a trace: only a few notes in the events log about its deletion. It seems a bit harsh to me to immediately start deleting instances on account suspension. A grace period would be nice. Surely there’s a grace period before deleting S3 objects, where RDS snapshots are stored?
It doesn’t matter, because RDS deletes all your automatic snapshots when you delete your instance. Or when they delete it, because your account was suspended for a few hours.
The end. Happy Friday.
Update: Amazon didn’t delete all my data when they suspended my account. I didn’t find a written policy guaranteeing this, but at least in my case, AWS automatically took a manual snapshot before terminating my RDS instance, and manual snapshots don’t get automatically deleted. I didn’t notice the automatic manual snapshot in the list with my manual manual snapshots from January.
It was a roller-coaster of a day, but I’m now happily back up with all the data I had before the suspension. Phew!