Operations | Monitoring | ITSM | DevOps | Cloud

Software systems are getting better at understanding themselves. The mix of richer telemetry, smarter pipelines, and agentic AI is shifting observability from a passive record of events into something more active and useful. That shift is what we mean by context engineering. We recently partnered with O’Reilly on a report by David Beale that introduces the discipline. Before you read it, here are five things worth knowing.

AI has changed how software gets built. It hasn't changed how software gets run. Most of the AI money in software has gone into the IDE: code generation, copilots, developer assistants, faster pull requests. That work matters. But writing software is one slice of the lifecycle. The harder problem, and the more expensive one, is running that software in production. Production is where systems fail in ways nobody predicted. Incidents don't stay inside one service.

PagerDuty fires. You open the incident. Title, timestamp, nothing else. Whatever context exists is in someone's head, in a Slack thread from two weeks ago, or in a runbook nobody has touched since the last reorg. We got tired of that. So we put an AURA agent behind a Slack bot and pointed it at our own production environment.

By: Tucker Callaway, CEO at Mezmo The market keeps saying “SaaS is dead.” That’s probably true, but it’s also incomplete. What’s actually dying is the idea that value lives inside a vendor-controlled black box. The next era is about utilities: unlimited coding capacity and unlimited analytical capability. And if those two utilities are real, then the vendor model has to change.

Builder in the loop is a Mezmo interview series focused on the engineers, product leaders, and operators shaping AURA, an open-source, MCP-native agent harness for production operations. This installment features Tony Rogers, whose work on AURA is less about building new features and more about trying to break them before users can.

Builder in the Loop is a Mezmo interview series focused on the engineers, product leaders, and operators shaping AURA, an open-source, MCP-native agent harness for production operations. The goal is to get past the polished product layer and talk through the decisions that matter when AI starts interacting with real systems. Key questions include: What should agents be allowed to do? How do they get better over time? Where should humans stay in the loop?

Single agents are a useful starting point for SRE workflows. They are not where the architecture should end. The first version is simple enough: connect an LLM to a few tools, give it a system prompt, and point it at your infrastructure. It can summarize an alert, pull logs, answer questions, and draft a useful next step. Then the workflow gets real. You add GitHub for runbooks, Kubernetes for cluster state, PagerDuty for incident context, Prometheus for metrics, and Mezmo for telemetry.

The first time an AI assistant suggests "restart the service" during a live incident and nobody on the bridge can tell whether that suggestion came from a current runbook, a stale wiki page, or thin air, you stop caring about model benchmarks. You start caring about what the agent actually knew, where that knowledge came from, and whether you can trust the chain of reasoning behind it.

An interview series with the people building Mezmo’s open-source agentic harness for production operations. Builder in the loop is a Mezmo interview series focused on the engineers, product leaders, and operators shaping AURA, our open-source, MCP-native agentic harness for production operations. The goal is to get past the polished product layer and talk through the decisions that matter when AI starts interacting with real systems. What should agents be allowed to do?

In a recent webinar, The Journey to Production AI, Andre Elizondo walked through what separates a working agent demo from an agent worth trusting on a 2 a.m. page. Live polls during the session put numbers behind a pattern most platform teams already feel. ‍ ‍ Most teams are early. The ones who are further along did not get there by shipping a flashier demo. They got there by treating production AI as a platform problem.