Operations | Monitoring | ITSM | DevOps | Cloud

Your AI agent just approved a return it shouldn't have, and nothing threw an error. Here's one agent request as a trace: the agent span, LLM calls, tool calls, and retrieval, with tokens and latency at every step. The fix wasn't the prompt. It was the data.

Most content comparing monitoring and observability stays at the definitional level: three pillars, metrics versus traces, known and unknown unknowns. This session skips that and looks at the operational side instead.

In Kepler ADE by GitKraken, you can make your own theme. Describe the look you want in plain words, and an agent builds it for you. How it works: Open Settings, go to Appearance, and click Create a theme Type the look you want. In this video I asked for gold and maroon The agent drafts the theme and checks every color for contrast before it saves. Then pick your new theme under Your themes Keep your theme for yourself, or share it with a pull request to gitkraken/kepler-themes so it can join the public gallery.

See who’s spending on AI, on which models, for what work, and what you got for it. Most companies can tell you what they spent on AI last month. Very few can tell you what it bought. In this demo, we show how AI Signals captures AI usage as it happens and ties every dollar to the person, team, and work behind it, across every model and provider.

As Datadog scaled to handle over a hundred trillion daily events, our engineering teams built their own custom query engines to keep things fast. While this helped teams move quickly, it also created silos that made it tough to analyze data across different products. To keep growing and support AI, we had to find a way to bring these engines together using open standards without slowing anyone down or risking reliability.

AI has become incredibly sophisticated and changed the way we build and power applications. But it has also introduced volatile complexity that challenges how we ensure reliability and protect user experiences. Join our panel of experts as they discuss the realities of building AI-powered applications. They’ll talk about how testing has evolved, share insights into how they do evals and experiments, and offer advice on how to get better visibility into how your AI is performing.

CircleCI automatically generates fix prompts for failed pipelines. Copy one into Claude Code, Codex, Cursor, or your coding agent of choice to diagnose failures, make fixes, and rerun pipelines with the CircleCI CLI.

Was this the worst week in cybersecurity? AI-discovered PHP2Shell vulnerabilities, the LiteLLM supply chain attack, and reports of unauthorized AI agents accessing bank accounts are exposing serious weaknesses in software security. But is this really a cybersecurity problem, or has AI exposed decades of poor software engineering decisions? In this episode of ShipTalk, Martin Reynolds and Adam Arellano break down a remarkable week in AI security and debate whether engineering or security teams should be responsible for fixing what comes next.

Learn how AI handles different types of testing, from simple unit tests to complex end-to-end setups. Discover where AI succeeds in writing test code and where human strategy is still needed. Learn more: speedscale.com.