Beyond Prompts: Securing and Observing Autonomous AI Agents

Cloud Native Night Munich

AI agents are no longer just generating text — they are opening pull requests, deploying to Kubernetes, and making autonomous decisions inside production systems. But how do we secure, govern, and observe systems that are inherently non-deterministic? In this meetup, we explore practical patterns for fine-grained authorization of AI-driven DevOps workflows and robust observability strategies for autonomous agents. Expect concrete architectures, cloud-native tooling, and real-world lessons for building agentic AI you can actually trust in production.

THIS MEETUP WILL BE HYBRID. | Zoom link will be available shortly before the meetup starts

*** AGENDA ***

• 6.30 PM | Doors open. Grab some snacks & drinks.
• 7.00 PM | Welcome & Intro – Start of hybrid meetup
• 7.15 PM | Beyond API Keys: Fine-Grained AI Agent Authorization for DevOps, Ankit Asthana (Cloud infrastructure Architect at SQUER), Tom Graupner (Head of Cloud & Platform Engineering as SQUER)
• ~8:15 PM | Mastering Agentic Observability: Reliable Multi-Agent Systems with OpenTelemetry, Nicolai Ommer (Software Architect at QAware)
• afterwards: More snacks & drinks, and a lot of time for networking with all attendees and speakers.

*** ABSTRACTS ***

Beyond API Keys: Fine-Grained AI Agent Authorization for DevOps, Ankit Asthana (Cloud infrastructure Architect at SQUER), Tom Graupner (Head of Cloud & Platform Engineering as SQUER)

AI agents are increasingly performing powerful DevOps actions like opening pull requests and deploying to Kubernetes. Yet most teams still rely on static tokens for authorization, which creates security and governance risks. In this talk, we present a practical pattern using OpenFGA and an Agent Gateway to enforce relationship-based, auditable authorization for AI-driven DevOps workflows—ensuring automation remains fast, controlled, and explainable.

Mastering Agentic Observability: Reliable Multi-Agent Systems with OpenTelemetry, Nicolai Ommer (Software Architect at QAware)

The software landscape is undergoing a paradigm shift. Agentic AI. While microservices operate deterministically, AI agents act probabilistically and often unpredictably, making them autonomous. This presents a significant challenge for platform engineers and DevOps teams. How can you monitor a service whose response varies with each call? How can you debug a 'hallucination' or an endless loop in decision-making?
In this talk, we will demonstrate why traditional 'golden signals' (latency, traffic and errors) are insufficient for AI agents and explain how they must be expanded. We treat agents as the new microservices and use the well-established OpenTelemetry (OTel) standard to shed light on the inner workings of LLMs and agentic workforces.

*****************

Please comply with our Code of Conduct.

Please note that photos are taken during the meetup. If you do not want to appear in the photos, please contact the meetup team at the beginning of the event.

Bitte halte dich an unseren Code of Conduct.

Während unserer Meetups werden Fotos aufgenommen. Falls du nicht auf den Fotos erscheinen möchtest, sprich bitte zu Beginn der Veranstaltung mit unserem Meetup Team.