Operations | Monitoring | ITSM | DevOps | Cloud

When we introduced Upsun Dispatch last week, we said we were building the platform layer for everything around the code. Today, you can apply to join as a founding design partner. Starting July 1, 2026, a number of engineering organizations will join us in prerelease. This is a selective, high-touch collaboration with teams who want to help shape what comes next. If you missed the introduction, you can catch up on Upsun Dispatch here.

At O11yCon, we talked to engineering teams across the industry, and the numbers are starting to get genuinely wild: Mixpanel DevOps Engineer Eddie Bracho told us their engineering team is generating 50% more PRs than before AI came into the mix (sorry). That kind of velocity is exciting, but it's also a pressure test for every part of your stack that isn't writing code, including your observability practice. Here's what we're hearing from customers about how that's playing out.

If you run Claude through LiteLLM, some of that spend is probably going uncounted – and you can’t see it, precisely because the data isn’t there. Routing through a gateway is messier than it looks: LiteLLM alone can carry Claude several ways – the OpenAI-compatible endpoint, and the Anthropic pass-through proxy that the native SDK and Claude Code use – and each path describes the same call differently.

A year ago, the sentence “I just deployed an app on GitHub” wouldn’t have made sense coming from me. I’m the VP of People at CloudZero; code deployments and I were not close friends. That’s changed. In this AI era, non-engineers are building, and I think that’s a genuinely good thing. But only if it’s tied to something that matters.

Many times, the hardest part of troubleshooting isn’t fixing the actual problem. It’s figuring out where to start. As engineers, it’s easy to lose count of how many times we’ve opened logs, then 10 metrics tabs, and another 10 tabs with trace queries, only to end up back in the logs trying to find a root cause.

Every layer of the modern network stack moves at cloud speed. If your connectivity doesn't, your entire strategy can stall. Co-authored by Fabio D’Avino This blog includes insights from Fabio D’Avino, a specialist in Network as a Service (NaaS) with more than seven years of experience researching, designing, and building global network services. Fabio’s work explores how organizations can modernize connectivity as cloud, hybrid, and AI-ready infrastructure strategies evolve.

When technology works, businesses thrive. Employees stay productive, customers stay connected, and critical services keep running. But when something goes wrong, the real challenge is not only detecting the issue. It is understanding what it affects, who may fell the impact, and how urgently the business needs to respond. That is the value behind the Kyoto release. The latest Skylar One update helps teams better connect IT health to business impact.

AI coding assistants like Claude, Cursor, Codex, GitHub Copilot have become standard tools in the modern engineering workflow. Developers use them to write code, generate tests, and review pull requests. But when something breaks in production, these assistants hit a wall: they have no access to your actual system state. They can reason about logs, traces, and metrics. They just can't see yours.

In our previous post on AI tokenomics, we looked at the rising cost challenge behind token-based AI systems. When enterprise IT teams rely on AI to reason through the same repeatable work over and over again, the costs to resolve those tasks may increase to an unreasonable level. That is where a deterministic IT automation platform becomes essential. A deterministic workflow follows predefined logic, meaning that given the same inputs and conditions, it produces the same expected result.

Knowing how to audit different types of IT hardware matters because a laptop, a server, and a network switch fail an audit for completely different reasons. Treating every device the same way during an audit means missing the checks that actually matter for each category, from disk encryption on an endpoint to firmware version on a router.