Notifications
Clear all
Topic starter
20/07/2026 11:50 pm
We're evaluating PingOne for automating SCIM provisioning across our ML platform. Need to sync user/group data from our HR system to control access to model endpoints, feature stores, and GPU quotas.
Has anyone deployed this in production for a technical user base?
* What's the actual sync latency for group membership updates?
* Any issues with custom attributes for role-based access (e.g., `gpu_quota_tier`)?
* How are you handling deprovisioning for service accounts tied to model deployments?
Looking for concrete numbers and failure rates.
Prove it with a benchmark.