Enable fan-out API orchestration to reliably handle large-volume Cloudability data requests

Large customers pulling cost, rightsizing, or container usage data across many accounts, dimensions, or date ranges often hit request size, timeout, or rate-limit constraints when calling the Cloudability API as a single request. Today, users have to manually split these into smaller calls, manage pagination and retries themselves, and stitch results back together client-side, which is slow, error-prone, and hard to scale as account counts and data volumes grow.

Native fan-out orchestration would let Cloudability automatically decompose a single large request (e.g., a cost report spanning hundreds of accounts or a long historical window) into a set of smaller sub-requests, execute them in parallel, and merge the results before returning a single consolidated response to the caller.

This would benefit FinOps teams and engineers running cost reporting at scale, particularly enterprises with large multi-account AWS, Azure, GCP, or OCI environments, as well as any tool or integration built on top of the Cloudability API (dashboards, chatbots, internal FinOps tooling) that currently has to implement custom batching logic.

Proposed behavior:

• Accept a single high-volume request (large account list, wide date range, or high-cardinality dimension set) without requiring the client to pre-split it.
• Internally partition the request into smaller sub-queries based on safe size/time thresholds, execute them concurrently against backend services, and apply retry/backoff on individual sub-request failures without failing the whole batch.
• Reassemble and return results as a single, correctly ordered/aggregated response (or via a job ID with polling, consistent with the existing queued-report pattern).
• Expose configurable concurrency limits and provide status/progress visibility for long-running fan-out jobs.