$ cat post/the-firewall-dropped-it-/-i-wrote-it-and-forgot-why-/-i-miss-that-old-term.md
the firewall dropped it / I wrote it and forgot why / I miss that old term
Title: When DevOps Met DORA: A Platform Engineer’s Perspective
April 11, 2022—It’s been a whirlwind month in tech. I’ve been thinking a lot about our platform and how we can stay ahead of the curve as these trends converge. Between AI/LLM infrastructure explosions and platform engineering becoming mainstream, it feels like every decision we make now could be pivotal.
The Day the Twitter Takeover Rumors Hit
It was April 12th—yesterday—as I sat in my office trying to debug a Kubernetes cluster, I couldn’t help but think about the craziness of the tech world. Elon Musk’s unsolicited $43 billion bid to take Twitter private had everyone talking. The drama was everywhere, even seeping into our developer Slack channels as people speculated about the impact on their favorite open-source projects.
Platform Engineering’s Growing Pains
On the platform side, we’ve been wrestling with how to manage our infrastructure better. With Kubernetes and containerization becoming de facto standards, it’s a challenge to keep everything running smoothly. We’ve seen a lot of excitement around serverless technologies like AWS Lambda, but also a bit of confusion on when to use them.
One recent argument in the team was about whether we should fully embrace WebAssembly for our backend services. Some argued that with the rise of AI and machine learning models, WASM could provide an efficient way to run complex computations at the edge. Others were skeptical, pointing out the complexity it introduces for deployment and debugging. I ended up on the side of skepticism—yet again.
FinOps and Cloud Cost Pressure
FinOps is another big topic that’s been weighing heavily on us. As cloud costs continue to rise, we’re under pressure to optimize our usage and find ways to reduce waste. One of my favorite tools for this is AWS Budgets and Cost Explorer. It helps us track spending and set limits on unexpected charges, but it can be a bit of a headache to configure properly.
DORA Metrics in Action
DORA metrics have become normalized across the industry. We’re now regularly tracking our lead time, change failure rate, deployment frequency, and mean time to recovery (MTTR). Last week, we had an outage that lasted over an hour, which pushed our MTTR way past our target. It was frustrating because it highlighted how much we still have to work on our incident response processes.
A Personal Lesson: Less Is More
One thing I’ve been thinking a lot about is the importance of simplicity in design and operations. The more complex our infrastructure becomes, the harder it is to manage and debug. In one recent project, we decided to remove unnecessary layers and streamline our deployment pipeline. It was hard at first, but it’s already paying off in terms of reduced complexity and faster deployments.
Looking Ahead
As I reflect on this month, it feels like every decision we make now matters. The tech landscape is changing rapidly, with AI/LLM models and serverless architectures pushing the boundaries of what’s possible. We need to stay agile and open-minded, but also grounded in best practices. Debugging the occasional Kubernetes cluster or figuring out how to use WebAssembly effectively are just part of the job.
For now, I’m focusing on making small, incremental improvements that add up over time. Whether it’s optimizing our FinOps practices, adopting DORA metrics more fully, or simply keeping things simple—every little bit helps.
That’s all for today. Stay tuned as we navigate this ever-evolving tech world together!
Brandon Camenisch
Platform Engineer & Engineering Manager