Operations | Monitoring | ITSM | DevOps | Cloud

The hidden cost of multicloud lives in operations and governance. Each provider develops its own console, runbooks, and audit posture, and the gaps widen quietly. The bill comes due at the next migration, audit, or incident.

When engineering teams evaluate AI agents, the first questions are usually about capability. Which model performs best? How much faster can it write code? What's the return on investment? Security, if it enters the conversation at all, tends to come later. Patrick Dawkins, Principal Software Engineer at Upsun, thinks that's backward. Over the past year, he's been building the infrastructure that enables AI agents to operate safely within engineering teams.

Migration projects fail in a predictable sequence. The technical work gets scoped. The timeline gets set. The engineering team starts moving workloads. Somewhere in the middle, dependencies surface that weren't in the original assessment, the double-run period extends beyond the budget allocated for it, and the project either stalls or completes at significantly higher cost than planned.

Every AI coding tool ships the same promise: your developers, faster. Autocomplete in the IDE, agents in the terminal, a working prototype before lunch. And it delivers, at least for the person holding the keyboard. The problem is that most of what it takes to ship software was never a solo activity, and that is the part the market keeps skipping.

Running workloads across AWS, Azure, and GCP is the operational reality for most enterprise engineering teams. The challenge isn't the providers themselves, it's what happens when each one accumulates its own delivery pipeline, its own security configuration, and its own environment management tooling. What starts as provider flexibility quietly becomes provider-specific complexity, multiplied across every team that ships.

Business continuity planning has a quiet assumption built into most of it: that the infrastructure the plan runs on will cooperate. Backup systems will be accessible. Recovery procedures will work as documented. The provider whose services underpin the critical path will be available, or at least recoverable within the window the plan specifies. That assumption is tested every time a major provider has an outage, a regional failure, or a service disruption that cascades across dependent workloads.

For the past two years, the conversation around AI in software engineering has focused on one thing: productivity. Engineers are shipping faster, writing more code, and completing work in hours that once took days. Every new model promises another leap forward. What gets far less attention is what all that speed demands from the people using it. Guillaume Moigneu, Field CTO at Upsun, has spent the past year watching engineering teams adapt to AI-assisted development.

If your team has adopted AI coding agents, you've probably noticed something strange: writing code stopped being the hard part. That's the shift Patrick, a principal engineer at Upsun, kept returning to in our latest Product Highlights conversation. He's spent twelve years here, most of them writing back-end APIs in Go, and the past year building with AI on our newest product, Upsun Dispatch. His verdict on where the bottleneck moved is blunt: "The code isn't really a problem anymore.