Self-Hosting: A Hobby or a Necessity?
The most expensive mistake of my career wasn’t a line of code; it was a “yes.” Years ago, when I said, “managing our own servers would be much more economical” for a project, I didn’t realize I was opening the door to a hobby. Since then, over twenty years in system architecture and operations, I’ve questioned the matter of self-hosting countless times. This post is here to share the real reasons behind this decision and my evolving perspective on it, without getting bogged down in technical details.
This isn’t just about cost optimization or a passion for technology; it’s also a story of seeking control, security, and even a bit of freedom. Managing your own infrastructure is a different world altogether, with its own responsibilities and rewards.
First Steps: The Desire for Control and Budget Pressure
It all started with the desire for more control. The limitations imposed by enterprise solutions, concerns about vendor lock-in, and of course, budget pressure, pushed me to set up my own servers. Back then, while working on a production ERP, I saw how critical it was to have full control over data flow and system performance. PostgreSQL database settings, Nginx reverse proxy configurations, even fine-tuning systemd units… all of these gave me the feeling of “I’m in charge here.”
The cost aspect, of course, couldn’t be ignored either. While cloud providers offered attractive services, bills could quickly escalate, especially for applications with high traffic or those running continuously. Managing your own hardware seemed like a more predictable and lower-cost option in the long run.
Realities and Updates: OOM and Other Nightmares
But over time, I realized that self-hosting wasn’t just about “self-sufficiency.” My first serious wake-up call came with a WAL bloat issue on a PostgreSQL server. The database logs filled up rapidly, slowing down the system, and I was woken up by an alert at three in the morning. Seeing the pg_wal directory reach 90% capacity triggered my initial panic. These kinds of problems rarely occur in infrastructures automatically managed by cloud providers.
Another example is my experience with containers. In a system I set up with Docker Compose, I suddenly saw a service exceed its memory limit and be terminated by the “Out Of Memory (OOM) killer.” I personally experienced how even soft settings for cgroup memory limits could sometimes be insufficient, and how momentary spikes could bring systems crashing down. Situations like these require not just writing code, but delving into the depths of the infrastructure.
Experiences like these show that self-hosting goes beyond just being a hobby. It’s a responsibility that requires deep system knowledge, debugging skills, and continuous learning.
When is it a Necessity, When is it a Hobby?
So, when does self-hosting become a necessity? For me, it’s directly related to the criticality of the application, data sensitivity, and the need for control. If the application you’re developing is the core system of a financial institution, an instant control software for a production line, or a platform processing sensitive personal data, the managed services offered by the cloud might sometimes be insufficient. In such cases, managing your own infrastructure allows you to control every step, from firewall policies to network segmentation.
On the other hand, self-hosting for a personal blog, a simple website, or a test project is usually part of a hobby or a learning process. In these situations, “managed” services offered by cloud providers can be much more efficient in terms of cost and time. Instead of dealing with your own servers, you can focus on your actual work.
In conclusion, self-hosting is a hobby; however, it is also an inevitable necessity under certain conditions. The important thing is to make this decision consciously and fully understand the responsibilities that come with it. Managing your own servers can provide you with immense power and knowledge, but you must be ready to pay the price for that power.
What do you think? Is managing your own infrastructure a passion for you, or just a cost-driven choice? Share with me in the comments.