I'm a one man DevOps team and my production runs for $0.64 a day. Here's the real story.

Аналитика рынка

Стек

Бесплатно, после входа в кабинет.

Откликнуться

Загрузите резюме — работодатели и рекрутеры из разных стран, которым нужен ваш профиль, смогут написать вам напрямую.

Описание

Most people writing cost posts here are spending their company's money. Mine comes out of my own pocket. I'm a solo DevOps engineer, I run the backend for a mobile app with real users, and the AWS bill is mine. That changes how you build things. Real setup, no dressing up. Two K3s nodes on t4g.small, one in each AZ. Two seperate database nodes, also t4g.small, running Postgres and Valkey. A tiny fck-nat box so I don't pay for a managed NAT. A bastion to get in. Backups go to S3, encrypted. Six small machines in total. I checked the AWS dashboard while it ran, it came to $0.64 a day. About $19 a month, everything included. Before building it I priced the managed way, the way the tutorials tell you. EKS is $73 a month just for the control plane. Small RDS is $13. Load balancer $16. Managed NAT $33, and that's before data. So around $135 a month in extras, more than my whole setup costs, times seven. I get why teams pay it, they are buying sleep. I wanted sleep too, I just couldnt pay that much for it. What broke first was not the cheap machines. It was my own monitoring. I installed the full Prometheus stack like everyone says, and it ate my 2GB nodes. Memory got so low, like 513MB free, that etcd couldnt reach consensus. The cluster was dying because of the thing that watches the cluster. I still laugh about that one. I ripped Prometheus and Grafana out, put in Uptime Kuma for health checks, it's like 100MB, and metrics-server for the basics. Lesson is simple, your monitori

Полное описание, компания и ссылка на вакансию — после входа. Войти

Контакты работодателя (email/phone/telegram) скрыты из публичного превью. Отправьте резюме — и работодатели смогут написать вам сами.