Cloud Resource Tagging
The question isn't whether your infrastructure will get complex, it's whether you'll be able to understand it when it does.
Cloud systems have gotten complex. With every "right tool for the job" decision, every microservice split, every managed service adoption, you're adding nodes to an increasingly tangled graph of infrastructure. That complexity isn't going away, it's the cost of building modern, scalable systems. The question isn't whether your infrastructure will get complex, it's whether you'll be able to understand it when it does.
Here's the thing: tagging is table stakes. It's not a nice-to-have or something you'll "get to eventually." Every untagged resource is a blind spot—invisible to cost analysis, impossible to correlate in monitoring, and orphaned from ownership. The sprawl of services, storage, compute, and data transfer across your cloud footprint demands machine-readable context. Tags provide that context. They're how you tell your tooling what something is, who owns it, and why it exists.
Good tagging unlocks everything downstream. Want to build cost dashboards by team or environment? Tags. Want monitors that alert the right owner when their service degrades? Tags. Want to correlate resource consumption to performance to spend for optimization decisions? Tags. DevOps, FinOps, SRE, all the practices we preach depend on the ability to slice and filter infrastructure programmatically. Without consistent tagging, you're flying blind and making decisions on vibes instead of data and the LLMs can't bail you out just yet.
You can't improve what you don't measure, and you can't measure what you can't identify. Visibility into the metrics that matter—cost per team, resource utilization by environment, ownership of orphaned resources—creates accountability. Accountability creates feedback loops. Those feedback loops expose inefficiencies that become growth opportunities for your engineers. A growing engineer spots an inefficiency and converts it into a solution. A growing engineering leader creates the visibility that makes those inefficiencies impossible to ignore.
The best time to start tagging was when you deployed your first resource. The next best time is right now. Treat it like compound interest, every tagged resource today makes tomorrow's governance, monitoring, and optimization easier. Start with the basics—environment, owner, cost center, service name—and enforce them at deployment. Your future self, staring at a billing anomaly at 2am, will thank you.
← all writing