Getting burned in the Cloud

It has finally happened to me… I’ve finally made a “mistake” in Azure that costed me money… In my defense the default setting changed, and I didn’t notice it.

The goal was to setup some backups in Azure with their cold storage. Pretty simple right? It’s super straightforward and I was up in running in 15 minutes later, starting to sign off for the night. The following week is vacation, and I don’t plan on doing a thing, I come home the following Sunday, decided to check Azure, and began sweating… I missed something? Something broke? Why was my Azure bill far higher than expected? I start to break down the cost and notice that RA-GRS (Read Access Geo Redundant) redundant storage was costing me money. Uhh what? I know for a fact I selected LRS (Local Redundant Storage). LRS was supposed to be my default selection? What happened? Let’s figure that out later, lets figure out how to stop the bleeding because these GRS data replication bandwidth charges are going to kill me.

Crack open the Azure documentation and start to look for my fix. Find it pretty quick, let’s just change the storage option. Great! Easy! Odd error: “failed to update storage account error sku update from standard_ragrs to standard lrs is not allowed at the moment”. Are you kidding me right now?? I begin searching, literally nothing online about this error… Turns out, certain SKUs become unavailable to change are certain times of the week…. Lesson learned. I legit now have to wait til the following day.

Monday comes along and I can finally make the change! Yes! Crisis averted! Close out for the day and come back Tuesday to make sure everything is cool… *insert buzzer noise* I’m still being charged for RA-GRS redundancy for our blob storage… Back to Azure docs… About 10 pages later, in very tiny fine print, it reads: “Cost of the original redundancy will continue until 30 days after initial change” Are you f%cking kidding me?! What load of BS is that?

At this point, I have to go nuclear, create a new storage account and move all of the data from one storage account to the other. Now we wait 6 hours for all of the data to move over, blow the old storage account away and call it quits. I read that doing these manual data transfers were free as long as it’s in the same zone. *insert buzzer noise again* WRONG, they don’t tell you that your write charges are going to go crazy… It cost more for this one-time data transfer than it did when we first started backing up all the data…

All is well in the world now, my mistake has been made, and I will live with this shame forever. But the show goes on. My key takeaways for this… Azure documentation is garbage. They’ll hide hidden costs in obscure documentation in tiny, hard to read print. As much as I’m a fan of Azure itself, I do find their documentation to be an abomination that burned to the ground.

Hopefully this helps the next poor soul who runs into a similar situation or prevents this situation from happening again. As always, feel free to reach out if you have questions! Don’t complain about the poor grammar, too much work.