Home / Companies / AuthZed / Blog / June 2025

June 2025 Summaries

2 posts from AuthZed

Filter
Month: Year:
Post Summaries Back to Blog
AuthZed has enhanced its cloud products by integrating with Datadog, thereby providing additional observability features that allow engineering teams to gain insights into authorization performance, latency, and anomalies without the need for extra custom tools. This integration enables customers to centralize observability data alongside other platform data in Datadog, facilitating event correlation across the entire platform. AuthZed continues to offer a web console with pre-built dashboards for infrastructure metrics, and now users can access similar dashboards in Datadog or customize their own. The integration simplifies monitoring for developers and SREs by offering metrics such as request latency and datastore performance, which are crucial for diagnosing issues and optimizing SpiceDB clusters. A streamlined setup process is provided for integrating AuthZed metrics into Datadog, affirming the convenience and efficiency of this new capability.
Jun 24, 2025 396 words in the original blog post.
AuthZed's Authorization Infrastructure for AI is designed to provide fine-grained permission systems for RAG pipelines and agentic AI systems, ensuring data protection, authorized actions, and compliance. It leverages SpiceDB, an open-source database inspired by Google Zanzibar, to store and compute permissions data, enabling scalable authorization controls for massive datasets. The infrastructure supports secure access control across every stage of the AI pipeline, from document ingestion to vector search to agent behavior, ensuring that sensitive customer or corporate data remains properly governed. It also addresses the authorization problem for AI by providing a better solution for teams building with LLMs, reducing the attack surface and mitigating risks such as exposure of customer records or unauthorized access to payroll data. With AuthZed's Authorization Infrastructure for AI, teams can safely scale their AI features without introducing security risks or violating data boundaries, ensuring that AI systems respect the same access controls as the rest of the system.
Jun 13, 2025 985 words in the original blog post.