The community member is experiencing timeouts with their database (weaviate) connections, resulting in errors when using the simple chat engine streaming method. They are looking for a way to gracefully retry the chat when this happens or any other recommended approaches to handle this issue. The comments suggest two potential solutions: try/except with a retry counter, or instantiating the client as a transient instead of a singleton. However, there is no explicitly marked answer.
We are having increase load on our application, and notice that sometimes our database (weaviate) connections times out. We are using the simple chat engine streaming method. This then sometimes result in an error like this: ''''HTTPSConnectionPool(host='xxxxxxxxx.europe-west3.gcp.weaviate.cloud', port=443): Read timed out. (read timeout=60)''' is there any way into gracefully retrying a chat when this happens? Or any other recommended ways in handling this?
haha yea fair, the case seemed to be that connections in the pool were dying, and that weaviate client does nothing in the way of refreshing/checking the connections. It's a bit hard to troubleshoot but our two options were indeed try/except with a retry counter or instantiating the client as a transient instead of a singleton
