News for hackers. Stuff that matters.
Re:HumanOps – Focusing on the human aspects of running infra... by alrs on Sunday June 26, 2016 @12:26PM (#11981421)
I would hope this boils down to "The end of on-call, forever."

Boeing doesn't require an aeronautical engineer in the cockpit of every 787, because they engineer their product to be operable. Have smart people build your infrastructure so that junior people can operate it, or else you wind up with unstable, burned-out, formerly-smart-people.

Re:HumanOps – Focusing on the human aspects of running infra... by mc32 on Sunday June 26, 2016 @03:43PM (#11982315)
Hear, hear!

I'd say, the blurring of work/off-work hours is one of the greatest coups of industry. Having someone on-call means they don't have to hire someone else to be there during "off-hours". I understand it adds costs --yes, pass them on [but someone else which cheap out and just have on-call and undersell us: pass laws to eliminate the choice.

I'm no fan of the industrial age with its "dehumanization"/abstraction of labor, but I do appreciate the hard delineation between work hours and non-work hours.

Re:HumanOps – Focusing on the human aspects of running infra... by specialist on Sunday June 26, 2016 @07:40PM (#11983384)
Not a coup. More like the continued ass kicking of labor by capital.

I had a leash (pager) in 1992 that was tugged often. The fur really flew if I didn't call back immediately. eg I almost got fired because I was in the shower at the gym (at 7:00am) and took 15 min to call back.

From what I can tell, since pretty much everyone I know has done a tour of duty at Amazon or equiv, things have only gotten worse.

"means they don't have to hire someone else"

Natch. But I see it more like someone else's failure to plan creates an (recurring) emergency for me.

Re:HumanOps – Focusing on the human aspects of running infra... by paulddraper on Sunday June 26, 2016 @02:20PM (#11981915)
You're kidding yourself if you think there's no one on-call to help with aircraft issues. The are in fact legions of aircraft controllers that fulfill the role of ops.

The only difference for a website is that no one but you and your customers care if it goes down, so it's up to you to provide the manpower (unless you outsource it to IaaS/PaaS providers).

Re:HumanOps – Focusing on the human aspects of running infra... by alrs on Sunday June 26, 2016 @02:44PM (#11982040)
Yes, 24-hour-industries need to staff 24-hours-a-day.

Note that air control towers are staffed with air traffic controllers, not aeronautical engineers. Also note that they're being paid to be there, and they don't ATC from their laptops in bed.

Re:HumanOps – Focusing on the human aspects of running infra... by paulddraper on Monday June 27, 2016 @12:09AM (#11984170)
> air control towers are staffed with air traffic controllers, not aeronautical engineers

Likewise, operations is different than development.

(Granted, you can have those roles fulfilled by the same people, but that's your choice to make air traffic aeronautical engineers. YMMV)

Re:HumanOps – Focusing on the human aspects of running infra... by alrs on Monday June 27, 2016 @12:52AM (#11984260)
If you're architecting systems and automating anything, you're in development.
Re:HumanOps – Focusing on the human aspects of running infra... by ams6110 on Sunday June 26, 2016 @08:39PM (#11983565)
Commercial aircraft can call their airline's maintenance department at any time for advice on problems that occur in-flight.
Re:HumanOps – Focusing on the human aspects of running infra... by protomyth on Sunday June 26, 2016 @02:26PM (#11981956)
Heck, speaking of transportation, there are people on-call for barges floating down rivers.
Re:HumanOps – Focusing on the human aspects of running infra... by web007 on Monday June 27, 2016 @12:36AM (#11984225)
Sort of yes, but mostly no.

Go read "Site Reliability Engineering" by Beyer et al. That's the clearest description for how on-call should work, and the way to get from Point A to Point B without burning anyone out. On-call is almost always necessaey, but should be thought of as high-deductible insurance. If you need it, use it, but just realize it's cheaper to prevent the problem than to fix it when things go wrong. And if you keep using it unnecessarily, your carrier will drop you.

Re:HumanOps – Focusing on the human aspects of running infra... by alrs on Monday June 27, 2016 @12:51AM (#11984253)
The first thing that jumped out at me in "Site Reliability Engineering" is that 1-2 incidents per on call shift is a reasonable amount. That's insane, unless you're only on-call twice a year.
Re:HumanOps – Focusing on the human aspects of running infra... by arkitaip on Sunday June 26, 2016 @03:57PM (#11982403)
That's a great example.