Core Insights - Datadog AI Research has launched two significant projects: Toto, an open-weights foundation model for observability, and BOOM, the largest public benchmark for observability metrics [1][2][4] Group 1: Datadog AI Research Initiatives - Toto is the first open-source foundation model focused on observability, trained on Datadog's internal telemetry metrics, achieving superior performance compared to existing time series foundation models (TSFMs) [2][3] - BOOM provides a comprehensive benchmark with 350 million observations across 2,807 real-world multivariate series, addressing unique challenges in production telemetry [4][5] Group 2: Technical Features and Benefits - Toto's zero-shot forecasting capability allows for immediate anomaly detection and capacity planning without the need for per-series tuning, which is essential for monitoring billions of ephemeral time series [3][5] - BOOM serves as an actively maintained resource for the research community, facilitating advancements in forecasting models specific to observability metrics [4][6] Group 3: Future Directions and Collaboration - Datadog AI Research aims to continuously release AI projects and collaborate with applied AI teams to develop tools that address customer challenges and enhance engineering workflows [5][6] - The open-source nature of Toto and BOOM invites contributions from the research and open-source software communities to advance observability forecasting [6]