Role notes

My role was a datacenter technician with some web development involved. As a technician at a bitcoin mining facility, my job was to ensure operations remained online and near a 100% uptime as much as possible. To accomplish this, there we monitoring tools we had to use such as Foreman and BTC tools as well as some of the software built inside the mining container housing units. However, the data we received in a visual form always came at a delay and lacking in information. We will touch base on that shortly. At this job, I had to learn how to work with machines from different manufacturers and how to diagnose and repair them. I began learning how to do PCB repairs to ensure our downtime when it came to hardware was minimal. Other repairs include control board and cooling fan replacements as well as PSU swaps. At this facility, we had to run power to all the containers, install all the racks, machines, etc. We started with empty containers when I was employed. By the time I left the job, there were nearly 2000 miners online and running. At times, we had to submit tickets for RMA processing. One of the biggest outcomes from this job for me was a side project I started. It is called ColonyWatch. It's an app designed to monitor your entire fleet of bitcoin miners. It supports many different types of mining machines and firmware. Managing our power throughout the day could sometimes prove difficult as some days it was hotter than others and certain areas of our operations had to be shut down. We did have to follow rules set by TVA to run on low power for several hours of the day. I grew used to working in a loud environment as the sound of whirring founds in the count of thousands can be quite loud! For the colonywatch project, I originally started off with just a simple NextJS app built to ping all the machines on a periodic basis. I noticed after 1000+ in-network commands sent and with that many replies coming in that perhaps nodejs was not the best suited for the task. My interface was kept simple using prebuilt UI libraries, slightly modified to fit my needs. Eventually, my backend was swapped over to rust using the help of GitHub Copilot and Claude Sonnet. before my departure there, I ended up developing, with the help of AI, a full monitoring solution that provided alerts via email and inhouse speaker system. We found that the system I wrote gave out notifications for downtime much faster than leading monitoring tools improving our overall uptime. My colonywatch project, at that time, lived within a docker container running on an older dell r610 I had sitting around and had upgraded.