AWS Builder Center
I Deployed to Production on a Friday. Here's Why I'll Never Do That Again.

I Deployed to Production on a Friday. Here's Why I'll Never Do That Again.

A relatable story about the classic "Friday deploy" mistake every developer eventually makes, what went wrong, and three simple habits that prevent it from happening again.

Cloud Data AI
Everyone warns you not to deploy on a Friday. I thought that rule was for other people.
The change was small — a one-line config update. Tested locally, looked fine, pushed it right before logging off for the weekend. Three hours later, at a dinner with zero laptop in sight, I got the alert. Something downstream broke in a way my local environment never would've caught.
Saturday became a debugging session instead of a day off. Here's what I took away from it:
1. "Small change" is not the same as "safe change"
The size of a diff has nothing to do with its blast radius. A one-line config change can take down more than a hundred-line feature update.
2. If you can't watch it, don't ship it
Deploying right before you're unreachable removes your ability to catch problems early, when they're cheap to fix. If you're about to be offline, that's exactly when not to push.
3. Monday-morning you will thank Friday-afternoon you for waiting
Nothing bad happens by holding a non-urgent change until Monday. Something bad very much can happen by rushing it out the door on a Friday.
I'm not saying never deploy on a Friday. I'm saying ask yourself: if this breaks, am I actually around to fix it? If the answer's no, it waits.
Who else has a "I thought I was the exception to the Friday deploy rule" story? Let's hear it.
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article