Old on prem server.
Up time 5 years (Basically the whole life of the server).
Rock solid performance.
One off cost.
New fancy cloud server.
2nd time in 3 months it's had to be rebooted.
Constant connectivity issues.
Constant monthly cost.
Randomly shits itself weekly.
Progress.
Up time 5 years (Basically the whole life of the server).
Rock solid performance.
One off cost.
New fancy cloud server.
2nd time in 3 months it's had to be rebooted.
Constant connectivity issues.
Constant monthly cost.
Randomly shits itself weekly.
Progress.
Comments
We’re in an era where a tiny number become insanely rich while the majority gets their noses rubbed into the shit pile they created again and again and .... before they start to comprehend the costs of being cheap and clueless.
Absolutely this.
Everything new that's introduced is an attempt to move responsibility onto some random process or third party.
"But I'm not paying for AWS. I'm paying for
Other than the mountain of problems with it, we've never had an issue.
The cloud definitely hasn't given me gray hairs.....
I need to check who is paying these features writers salaries.
Ouch.
Uptime of 4 or 5 years.
6 folks doing weeks of planning- move day the box never powered back up.
Thankfully we had a plan B…but what stress…
Typical app that exploded with popularity, so couldn’t move 😳
I just go by how much more my work is interrupted by problems now and whatever the sysadmin makes up.
On prem server goes wrong, you fix it 🫵
Cloud server goes wrong, raise a ticket 🤷♂️
Ephemeral workloads are the only way to fly.
3-4k a month, as sold to the brass, is currently nearer 15-16.
Oddly the thought is. Wow that guy is a hero for sorting out these problems and not wow that guy needs to be fired.
People confuse me.
"Oh it's cloud so we can just spin up a new server whenever we want"
And
"Our server config is so complex it's going to take two months to set up a new one"
Depending on if he's talking to a manager or someone technical. 😄
The principle of K8S, for example, is brilliant. But the technical complexity 🤦🏻. You need other complex tools (helm, argocd, for example) to manage the complexity 🤯.
they were, how do you put it, fucked.
But that aside, the problem isn't the cloud, it's trying to apply on-prem paradigms to it. Usually because the admin doesn't know any better. Sorry.
Ours is a batch system with a more or less completely predictable load.
The fact it was sold to us on quick scalability and management lapped it up makes me fear for my sanity.
I left after a disagreement with Director #1, but Director #2 was still on speaking terms with me on the down-low. He shared me a few screenshots...
Easiest 3 month contract ever just undoing what the clownshoes that came after me had done. It was Docker Swarm + 2x 128 core boxes doing batch work + a handful of EC2 running the frontend.
He came in and immediately ditched my project and hired a bunch of bad hires.
Also managed AWS Kubernetes lmao. Of course it is.
(K8s I'm 90% sure is also a make-work project used by Google to keep their competitions wheels spinning but going nowhere. But that's a different rant.)
(either its heinous code that's unreadable and has lots of side-effects that are non-obvious or that someone is just coding by prompting copilot to do their job for them)
Someone touting that they have vendor certs is a 'smell'.
Good luck