Memory Batch Ingestion Concurrency Limit
To prevent server overload, Hana caps the number of memory batch ingestions a single user can run at the same time. When you reach this threshold, new ingestion requests are rejected until one of the active ingestions completes. This ensures faster, more predictable ingestion for everyone.
The current limit is 3 concurrent memory batch ingestions.
How the limit is applied
- Scope: per user, per active organization
- Counted as in-progress: batches with ingestion status
IN_PROGRESS - Not counted: batches with
autoResyncEnabled = true(auto-resync worker runs are excluded) - New ingestion/sync requests above the limit are rejected with a
429-style rate-limit response
Tips
- Queue large sources during off‑hours to finish sooner.
- Group related docs into a single batch where possible.
- Start with a small test batch to verify access and formatting before scaling up.
Example
If you start a fourth ingestion while three IN_PROGRESS ingestions already exist for your user in the same organization, the new request is rejected. Wait until at least one in-progress batch transitions out of IN_PROGRESS (for example INGESTED, ERROR, or ABORTED) before starting another.