Data platform engineers, analytics architects, and Databricks administrators who need to give cloud-based Databricks workloads live query access to on-premises AIStor data, and who are responsible for configuring, securing, and monitoring that sharing infrastructure in production.
.svg)
The open sharing endpoint must be explicitly configured before any tokens are created, as AIStor embeds it into every pre-signed URL and profile.share file at token creation time. An empty endpoint causes Databricks to fall back to an internally detected address that external consumers cannot reach.
A single AIStor Table Share endpoint serves both Iceberg and Delta tables simultaneously from the same share and token, with Databricks executing cross-format SQL joins across both table types without any format-specific configuration on the consumer side.
All six production settings (endpoint, pre-signed URL expiry, per-token rate limit, global rate limit, OAuth toggle, and OAuth token expiry) have corresponding Prometheus metrics available on the v3 endpoint, enabling real-time observability over request volume, response time, and rate limit saturation as consumer load grows.
.svg)
