MyKE / About me

Created Sat, 24 Aug 2013 15:19:09 +0000
741 Words

About me

Hi, I’m Michael Sabrnak aka MyKE - a Staff Site Reliability Engineer at SolarWinds based in the Czech Republic, with 12+ years of hands-on experience across infrastructure, cloud, and platform engineering.

What I do

I build and maintain the platforms that keep engineering teams productive. My day-to-day revolves around Kubernetes (AWS EKS & on-prem), infrastructure-as-code (Terraform, Ansible), CI/CD pipelines (GitHub Actions, TeamCity, CircleCI), and observability - making sure things run, scale, and recover gracefully.

At SolarWinds I’m part of the internal EngOps / SRE team where we practice “drink your own champagne” - we rely on SolarWinds Observability (SWO) to monitor the very infrastructure we build. I’ve worked on projects ranging from managing Kubernetes clusters at scale, building self-service developer tooling, to designing alerting strategies that cut through the noise.

My journey

I started in IT back in 2013 as a computer engineer doing SMB outsourcing, then moved into system administration at VSB - Technical University of Ostrava where I managed Active Directory, Exchange, SQL clusters, Citrix, and VMware environments. From there I joined Tieto (now TietoEVRY) as a Windows Senior Technical Specialist and later Infrastructure Architect - automating everything I could touch with PowerShell and designing customer-facing infrastructure changes.

In 2019 I joined SolarWinds as a System Engineer and progressively moved into the SRE world: Site Reliability Engineer - Senior SRE - Staff SRE (since October 2023). Along the way the stack evolved from pure Windows Server into a hybrid world of Linux, containers, cloud, and GitOps - and I’ve enjoyed every step of that transition.

Speaking & community

In April 2024 I co-presented at THWACKcamp 2024 - SolarWinds’ annual virtual IT learning event - with a session called “It’s a Beautiful Day in the SWOC-erhood: Real World Applications of SolarWinds Observability”, showing how our EngOps team uses SWO for real, not just as a demo.

Voyscout and the AI SDLC

In the summer of 2026 I built Voyscout - a flight and holiday price tracker for Czechia and Central Europe, and my first bigger side project after a long time. It is primarily vibe-coded: AI agents wrote almost all of the code, while I put my SRE and AI SDLC experience into everything around them - an agent harness with codified skills, hooks and allowlists, separate agent seats (implementer, monitor, supervisor), a review ladder that starts with a local model on my own GPU, the same CI gates for agents as for humans, staged autonomy with structural guardrails, and monitoring of both the product and the agents themselves (deterministic sensors, a watchdog, Netdata).

Today the project mostly lives its own life and I manage it from a distance. I wrote about how it works in AI SDLC road to Voyscout, and about the product itself in Voyscout - my flight and holiday price tracker. If you want to run AI agents safely yourself, my Claude Code Sandbox template is a good place to start.

Beyond work

Outside of the office I’m a hands-on tinkerer. I run a Home Assistant setup with Zigbee devices, manage my home network on UniFi gear (previously MikroTik), play with Flipper Zero and hardware security, and keep a small fleet of Raspberry Pis running various services. I’m also a satellite TV hobbyist (Enigma2 / OScam) and I care about privacy - I self-host what I can and use tools like SimpleLogin to keep my digital footprint in check.

I’ve also been running Erebor - a Czech Ultima Online private server with a Lord of the Rings theme - from my home infrastructure since 2015. It’s a niche community for a classic game, but keeping it alive for 10+ years has been a fun long-running project. Behind the scenes it’s a full production setup - source control, automated deployments, centralized logging with ELK, a community wiki, and secure remote access for the admin team. I also maintain a UO tools archive with downloads and guides from the early days.

This blog

I’ve been writing here since 2013. The blog started as a personal knowledge base - quick fixes, troubleshooting notes, and scripts I needed to reference again. Over time it grew into 90+ posts across 43 categories, covering everything from Windows Server internals to Kubernetes networking to Flipper Zero hacks. The common thread: real problems, real solutions, no fluff.

This site is powered by Hugo and hosted on Cloudflare Pages.

Find me elsewhere