One of our clients is performing on-demand extraction using PriceHistoryExtractionRequest. They are experiencing significant latency increases under parallel loads, so we would like to confirm extraction limits.
Client usage:
- Multiple application instances share a single DSS user account.
- Each instance executes the PriceHistoryExtractionRequest on demand via ExtractWithNotes.
- During peak periods, approximately 10-40 extraction requests are sent simultaneously.
- Regarding the limit on concurrent extraction processing: The Best Practices document (ver. 9.2) states that the "Concurrent extraction processing limit is 2 per report template per user." Does this limit also apply to PriceHistoryExtractionRequests? Specifically, if 10 PriceHistoryExtractionRequests are submitted simultaneously from the same user account, will only 2 be processed and the rest be queued?
2. Regarding queue behavior: Are requests that exceed the concurrent processing limit automatically queued? Is there a maximum queue size or timeout?
3. Regarding the rate limit for on-demand extraction: Is the Fair Usage Limit (30 on-demand extraction requests/60 seconds) mentioned in the documentation a per-user limit? Am I correct in understanding that HTTP 429 is returned when the limit is exceeded?
4. Regarding recommended configuration: In our use case (extracting multiple time-series stock price data for different stocks in parallel), please let us know if there is a recommended approach to maximize throughput within the fair usage range.
5. Regarding limit increases: Is it possible to increase the limit of concurrent extraction processes for this client’s account?
Regards,
Hiroko
