If you've got a good automated deployment process and confidence in your tests, go for it.
At a previous team, we deployed M-Th in the mornings. We could have deployed on Fridays but we opted for "send your big changes to the Release Manager on Thursday so it gets two days of vetting before the Monday release."
I started hacking on IRC scripts in the late nineties and have been hacking on computers (and businesses) ever since. I love to build, learn, and share.
Good points. We have both of those. The thing we've internalized as a tiny team operating a high-throughput, high-availability app is that even with the greatest confidence, when things do break, it's with bad timing.
Deploying on the last day of the week is usually NBD, but I'm risk-averse, so I'd rather deploy big changes when I don't have to give up my evening/weekend to support them. I do that calculation whenever I deploy something major (we deploy small changes all the time).
I started hacking on IRC scripts in the late nineties and have been hacking on computers (and businesses) ever since. I love to build, learn, and share.
I've been working on this little side project... It's a Rails engine/gem that automates email sequences to users like Intercom.com does, but it's self-hosted. It's finally ready to actually replace the user-onboarding emails that we were sending through Intercom. I had been planning to roll this out for a few weeks but kept getting sidetracked... This morning I decided that it was happening today no matter what. :)
¯_(ツ)_/¯
If you've got a good automated deployment process and confidence in your tests, go for it.
At a previous team, we deployed M-Th in the mornings. We could have deployed on Fridays but we opted for "send your big changes to the Release Manager on Thursday so it gets two days of vetting before the Monday release."
Also, no disrespect for people who choose not to. You do you.
Good points. We have both of those. The thing we've internalized as a tiny team operating a high-throughput, high-availability app is that even with the greatest confidence, when things do break, it's with bad timing.
Deploying on the last day of the week is usually NBD, but I'm risk-averse, so I'd rather deploy big changes when I don't have to give up my evening/weekend to support them. I do that calculation whenever I deploy something major (we deploy small changes all the time).
For sure. I think that's a perfectly valid reason for not deploying on a Friday or even right before 5 on any other day.
So since you are deploying today, what was so important? :)
I've been working on this little side project... It's a Rails engine/gem that automates email sequences to users like Intercom.com does, but it's self-hosted. It's finally ready to actually replace the user-onboarding emails that we were sending through Intercom. I had been planning to roll this out for a few weeks but kept getting sidetracked... This morning I decided that it was happening today no matter what. :)
So far, so good. 👍
Very cool!