Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

if you have the monitoring in place, and you have done proper canarying, friday evenings are not terrible


Of course. You're more than welcome to risk your own weekend over trivialities. The rest of us only ask that you please keep us out of your plans.


You never were. And there was a reason for that. Maybe you are not the person to put his 150% on the line when needed. It is ok. You do not have to be. Then you get paid accordingly.

You also didnt ask what could prompt before weekend pushes. It could be money, reputation or or compliance. Certain projects are out of your control and you have to implement in weeks notice, or have a deadline imposed by laws. Some are losing revenue so each day you do not release you are bleeding money.

So please, by all means, take your weekend off, while i plan for worst cases and make it so that friday releases are "trivial and safe" to deploy. I do not have to use it, but if i do, i do not need you around, it will be mostly automated anyway.


The problem isn't the day of the week but the fact that it is the day before time off.

If you work weekends then by all means push on Friday evening. However, don't push anything the before your next vacations.


and if you have someone else on call, Friday evenings are perfectly fine

:p


If and when this happens, the pusher person usually is the oncall.


monitoring only helps detecting a problem. and even with canarying there could still be a bad thing in production. it happened everywhere, even the big ones had akward problems. it's human to make errors, so if you do not have luck such a problem could occur on friday evening/night.


There is no doubt, but you see, there are legitimate reasons to do it.


yeah well I always do releases on friday and we have monitoring/canary. but sometimes a minor/major bug slips trough, but our customers only work from monday-friday and so it's not that big of a deal. and it barly happens and after that we add a regression test so that at least the same error won't happen again. funnily most errors on our codebase happened in client side or sanitazion problems. and the client side errors go away once we added all the types to typescript (we converted js to typescript and added any, so we still need to "correct" like 40% of our code base)


That is kind of my point. I believe over time, the monitoring gets solid enough that you can do abrelease anytime. I heard often teams not wanting to do a release, but i try to go towards "i can measure a good amount of metrics and slices that are important and can measure their impact in a controlled manner" which helps for friday releases, or any other release for that matter.

I dont do friday releases often, but if it comes to it, it is no brainer. I approach it pragmatically, and this helps people.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: