Pyarango connection management

[Leena Bahulekar]

Hi,

I am planning to use pyarango driver to connect to the ArangoDB. I have a pyramid web application which is supposed to provide CRUD operations on the knowledge graph. I had some specific questions around the connection management:

1. https://www.arangodb.com/docs/stable/http/general.html - Documentation states that the client can implement connection Pooling since TCP connections are expensive. I am looking through the code for pyarango but do not see a connection pool implemented. Please guide me to it if I am missing something. GO/ Java drivers seem to document that they have a connection pool implemented

2. We are planning to create a new pyarango connection to arangoDB per HTTP request in the pyramid web application. If there are several pyarango calls made in the same request flow, I understand that they will be made on the same pyarango connection -TCP connection. Thus saving some delays. 

However is this a good practise, especially since I do not see connection pooling implemented in pyarango. Or should my pyramid web application implement a pool of pyarango connection objects? 

3. I do not see any connection close equivalent code in the latest pyarango code. Does that mean that the connection - in my case - would eventually be closed only after TCP connection timeout?

Any help is greatly appreciated.

Regards,

Leena.