January 13, 2020. The year was just starting to unfurl its complexity, and I was in the thick of a lot of it. Platform engineering was finally taking shape, and the tools like Backstage were slowly gaining traction. But that wasn’t all; SRE roles were becoming more common, Kubernetes was showing its complexities, and eBPF was starting to gain some attention. It was also around this time that we began grappling with the reality of a remote-first workforce due to looming global events.

The Rise of Internal Developer Portals

One of the projects I was heavily involved in during 2019 was the internal developer portal, built on Backstage. We were excited about the promise it held for streamlining our engineering processes and improving collaboration within the team. However, building something that feels like an internal product can be tricky when you’re still figuring out how to manage a growing number of services and dependencies.

Backstage promised a lot: a single source of truth for all your infrastructure needs. But with every feature we added, it felt like there was more work to do. We had to balance the desire for a seamless developer experience against the need for maintainability. It was a constant battle between complexity and simplicity.

Kubernetes Complexity Fatigue

Kubernetes, our trusty friend in 2019, began showing its growing pains. As we added more services to our cluster, managing them became increasingly complex. Deployments, updates, and rollbacks were becoming a full-time job. The initial excitement of having everything run on the same platform was waning as the reality of ops hit us.

I remember arguing with my team about whether it made sense to invest in an advanced Kubernetes setup or stick with simpler solutions that might be more maintainable but less flexible. In retrospect, I think we took a few missteps along the way. We should have considered smaller, more targeted clusters for different parts of our architecture instead of one giant monolith.

Remote Work and Infrastructure Scaling

Another trend that was quietly building steam in 2019 was the shift towards remote work. As it turned out, by mid-2020, we found ourselves scrambling to support a fully distributed workforce. Our infrastructure needed to scale in ways we hadn’t anticipated. We invested heavily in tools like ArgoCD and Flux to manage our GitOps processes, but there were still plenty of rough edges.

One particular night stands out in my mind as a turning point. I was up late trying to debug an issue with one of our Kubernetes clusters when the power went out for half the office. Our data center had some old infrastructure that couldn’t handle the load. It was a stark reminder of how brittle our setup could be, and how much more work we needed to do to make everything resilient.

Personal Reflections

As 2019 drew to a close, I found myself reflecting on what had changed in tech over the past few years. Backstage and other platform engineering tools seemed promising, but they came with their own set of challenges. The proliferation of SRE roles was a sign that our ops teams were becoming more sophisticated and essential.

On a personal level, I realized how much I’d grown as an engineer. I had faced some tough decisions and learned valuable lessons about trade-offs between complexity and simplicity. It wasn’t always easy, but it was rewarding to see the impact we could have by making difficult choices and pushing ourselves to do better.

Looking back, 2019 felt like a transitional year in tech. We were building out these new tools and processes while also dealing with the reality of change—whether that meant shifting work environments or grappling with the complexities of modern infrastructure.

Stay tuned as I continue to navigate this exciting yet challenging landscape in 2020.