The machines responsible for ETL run independently, and can number in the hundreds. It has been difficult, or impossible, to apply a global request rate limit to other services it uses.
I believe the only solution is relay all requests through a central server that can apply rate limits.
The machines responsible for ETL run independently, and can number in the hundreds. It has been difficult, or impossible, to apply a global request rate limit to other services it uses.
I believe the only solution is relay all requests through a central server that can apply rate limits.