Version 0.1.0-beta.8

TL;DR: Write directly to catalogs (create or replace a single item, delete one, or apply a batch of adds, updates, and removes in one call). Failed scenario, report, and catalog writes now explain what went wrong with a short, consistent reason instead of an opaque upstream error. The global rate limit is higher, and organization listing and customer identifier labeling are more reliable.

Write catalog items directly

• Create or replace a single item in a catalog with upsert_catalog_item. Pass the item's fields as a flat map of field name to value. This is a full replacement, so include every field you want the item to have, not just the ones that changed.

• Delete a single item with delete_catalog_item. Deleting an item that's already gone counts as a success, not an error.

• Apply a batch of adds, replacements, and removals to a catalog in one call with import_catalog_items.

upsert_catalog_item and delete_catalog_item work against both catalog generations automatically. import_catalog_items is v2 only — use the single-item tools one at a time for a v1 catalog. A v1 write is queued in the background; a v2 write returns a job ID you can poll with list_catalog_jobs before treating catalog search results as current. Catalogs backed by a Data hub item collection don't accept these writes — use the Data hub item collection tools for those instead.

Clearer errors from failed writes

Scenario writes, report calculations, and catalog writes now return a short, classified reason when they fail, so you can tell an upstream outage apart from a rejected request at a glance. Validation failures still include the specific field and message the backend rejected; outages and authentication failures no longer echo the raw upstream response.

Higher global rate limit

The global rate limit for read and write calls increased from one call per second to three; the per-minute ceiling of 60 is unchanged. Independent calls can now run in parallel without hitting an immediate rate-limit error. See performance and rate limits.

Bug fixes

• Listing customers could mislabel identifier values. For example, an email address showing up under an unrelated identifier name. Identifiers are now labeled correctly.

• Listing your cloud organizations could fail on a brief upstream delay. It's now more resilient to those delays, and a listing that still can't complete returns a clear prompt to retry instead of failing outright.