Uploaded September 2026 | Updated November 2026, 2 days ago
Ray enables large distributed training workloads across GPU fleets, but understanding what happens inside each job remains difficult.
At Ray Summit 2026, Marina Petzel and Ashley Chu from Datadog share how they built a unified job-level view that preserves history beyond cluster lifetimes, traces work across Ray processes, and connects failures with GPU health, utilization, and cost.
You'll leave seeing why distributed AI observability should center the job, not just the infrastructure.
Liked this video? Check out other Ray Summit breakout session recordings: youtube.com/playlist?list=PLZNaJyYZRPdo
Subscribe to our YouTube channel to stay up-to-date on the future of AI: youtube.com/c/anyscale
đź”— Connect with us:
LinkedIn: linkedin.com/company/joinanyscale
X: https://x.com/anyscalecompute
Website: anyscale.com
Ray enables large distributed training workloads across GPU fleets, but understanding what happens inside each job remains difficult.
At Ray Summit 2026, Marina Petzel and Ashley Chu from Datadog share how they built a unified job-level view that preserves history beyond cluster lifetimes, traces work across Ray processes, and connects failures with GPU health, utilization, and cost.
You'll leave seeing why distributed AI observability should center the job, not just the infrastructure.
Liked this video? Check out other Ray Summit breakout session recordings: youtube.com/playlist?list=PLZNaJyYZRPdo
Subscribe to our YouTube channel to stay up-to-date on the future of AI: youtube.com/c/anyscale
đź”— Connect with us:
LinkedIn: linkedin.com/company/joinanyscale
X: https://x.com/anyscalecompute
Website: anyscale.com










