3 ms·
Having worked in AWS DCO- there's a lot of factors. You were under a shit manager. APAC Regions manager is inept. Dublin was also filled with fucking idiots.
by sdmetacarpus 10y ago
Having worked in AWS DCO- there's a lot of factors.
You were under a shit manager. APAC Regions manager is inept. Dublin was also filled with fucking idiots.
DCO's upper management in general is a fuckfest of people who want to play big man but have no idea what metrics ARE. They have no basic grasp of math nor do they know how to find what they need. I'd be shocked if they could tell you what an average value is from a set of numbers.
That said- there was a huge fucking technical gulf between the average DT and an engineer. So many DTs wanted to be SDEs or Sysengs, but their real world technical competency is that of a junior sysadmin at best so they saw no real room for promotion.
These days, automation removes the learning part of it. AWS back in the day- you could have techs that would qualify as mid level sysadmins at other companies. Today- the scale is so huge and the fleet so automated that it's mostly just throwing motherboards at problems and swapping harddrives. They want to turn the datacenter work into the equivalent of warehouse work so that they can hire any body off the street can do the job. The hilarious irony is that you now get former aws datatechs with it on their resume entering the technical market who have no business doing so.
Finally- I think it's an issue of just large companies in general. As the group grows, you inevitably hire assholes who make asshole moves and poison the corporate environment. Amazon grew so fast and hired so fast- there's no way that this could be avoided.
DCO just kinda sucked if you were less than a level5.
- jsudhams 10y agoif this is real i see they will have issues in future they charge very well for the service they provide so i assume they should have good engineers stick to them for long term growth.
- jon-wood 10y agoI think the point being made is that their automation is at a point where they don't need good engineers on data centre ops. Anything which requires skill to diagnose and fix is software defined and managed remotely, which leaves nothing much more than racking servers and swapping out the hard disk with a blinking light on its caddy.
- ismail-khan 10y agoThis is really great insider information, thanks for sharing!