404 now. New link is https://engineering.shopify.com/122596804-why-shopify-moved-to-the-production-engineering-model
One of the comments says:
I’ve heard [it] said repeatedly that if you try and put the developers on call you’ll have a hiring problem, the ‘best’ people won’t want to do that (and have a choice of employers)
Do people here count it as a negative against a potential employer? Seems to me to be a logical part of my responsibilities (and accounted for when determining compensation).
I’ve always found it extremely difficult to convey how stressful being on call can be, regardless of whether one gets paged or not. Compensation really needs to take that into consideration when the team size on rotation is <= 5.
I’m less willing now I’ve got a toddler waking me regularly.
I prefer an on-call rotation to the alternative which, I believe, is that your more senior devs are always on-call. I view the rotation as a gift. I have time that I do not need to worry about having my phone with me or what is going on with the platform.
As an aside, I think it also ensures that issues get fixed. If you are paged on the same issue repeatedly then there is an opportunity for mitigation or automation.
Absolutely, if it’s presented as “we don’t have ops, everyone just does rotation” it falls deep into the immature management category. Not making a plan for executing the product after you ship it is just insane amounts of immaturity.
If there is a core ops team, and a small number of devs rotate in helping out over time.. it’s not so bad if it’s a brand new product, and you’re trying to figure out how and where it fails. I say new product as many features and deep product issues haven’t been worked out yet thus need development to investigate quickly.
If there is a core ops team, and a lot of devs rotating in and out, but it’s a product that’s been shipping for over a year.. then we’re back to poor management. If your engineering team (inc mgmt) can’t create a concrete plan around shipping a product, and operating the product with your customers – then you’re probably not in a good place.
Huge caveat: this does not mean the company won’t be successful. Amazon is a great example of somewhere with horrifying management, incredible turn over, shit quality of life – but they make money hand over fist. Deciding if it’s a place you want to work is another matter.
I used to work somewhere where the on-call policy was: rotate within team for week-long chunks, and you have a monetary bonus when you are on call. In my case, I ended up on call every other week, so it upped my salary by an okay bit.
How much was the bonus?
At Google we do the same thing and the bonus is 33% of your salary.
I count it as a negative against a potential position. I have turned down a job that offered substantially higher salary than the one I took, and where the interviewer was possibly the best programmer I’ve met, because it involved being on-call. Uninterrupted sleep is something I value very highly, and at the time I didn’t quite trust even my own code to never have problems.
absolutely. if I have a choice of jobs, or teams/roles within a job, I’ll always take the one that doesn’t involve being on call