I'm replicating a dataset (60k records) with 100 threads (Python's builtin ThreadExecutor) at a time. And I find that from time to time I keep getting these warnings from the 'requests' library: "Connection pool is full, discarding connection: xxxxxxxx.algolia.net".
The ability to edit the _poolmaxsize configuration variable when initializing the search client would solve this kind of issues, perhaps even allowing to edit other requests configurations.
Steps To Reproduce
Initialize a ThreadExecutor of 100 max workers.
Create a function that uses 'save_object' to store a record in a predefined index.
Loop through thousands of records, for each record:
-Add the function that saves the record with the record data to the executor
Description
I'm replicating a dataset (60k records) with 100 threads (Python's builtin ThreadExecutor) at a time. And I find that from time to time I keep getting these warnings from the 'requests' library: "Connection pool is full, discarding connection: xxxxxxxx.algolia.net".
The ability to edit the _poolmaxsize configuration variable when initializing the search client would solve this kind of issues, perhaps even allowing to edit other requests configurations.
Steps To Reproduce
Initialize a ThreadExecutor of 100 max workers. Create a function that uses 'save_object' to store a record in a predefined index. Loop through thousands of records, for each record: -Add the function that saves the record with the record data to the executor
You should get the warning every ~1000 records.