Last Tuesday at 06:00 AM, when the Pipeline-health monitor sent a 'DEGRADED' email, I knew exactly what was coming. It was the silent scream of the systems I built and manage myself. The price of building my own ecosystem as an indie hacker means being on call 24/7. In this post, I'll share how I experience the on-call responsibility for my own systems, what happened to me, and what I learned in the process.
Most people think of on-call as a task rotated among teams within the complex infrastructures of large companies. For me, the situation is a bit different; on-call is a continuous conversation between me and my VPS, which hosts my 13+ Docker containers. This conversation usually starts with the sound of an alarm at midnight or while I'm having breakfast on a weekend.
This situation is more than just a technical task; it's also a mental burden. Living with the thought that something could go wrong at any moment affects my productivity and even the quality of my rest. On the path I've taken to develop my own products, I have to be constantly vigilant to avoid being crushed by the operational load of these products at some point.
The Realities of Being On-Call for Yourself
The phrase "a prisoner of my own system" might sound dramatic, but I've clearly understood the reality of this situation through the events I've experienced. For me, on-call isn't just about hearing a 'pager' sound; it's about instantly identifying why that pager is ringing and reaching a solution. Some scenarios I've encountered during this process highlight the seriousness of the situation. This carries a different meaning, especially for an indie hacker carrying this burden alone.
Once, when I experienced resource exhaustion on my own VPS, I saw the kcompactd process in the system using 92% CPU. This meant the kernel was completely locked up while trying to optimize memory. Since even sshd couldn't accept connections, managing the system remotely became impossible.
Top comments (0)