Fix memory leak on large responses

[mostafa]

A table with 158369 rows and the response size of 67735631 bytes causes subsequent queries to hang. This could be related to #175.

Observations

1. The query, select * from table1;, causes the client to get stuck and doesn't return any meaningful results, while the same query with LIMIT will work, that is select * from table1 limit 158369;.
2. If the limit is set to 7300 or less, the result is cached correctly and later read from the cache, but if it exceeds that amount, the cache plugin fails to receive the data, hence it doesn't caching anything and returns this errors. This is related to Reading large inputs from server causes plugins and Run to hang on onTrafficFromServer hook #175.

   2023-03-09T17:44:35+01:00 ERR Hook returned an error error="rpc error: code = ResourceExhausted desc = grpc: received message larger than max (4492364 vs. 4194304)" hookName=HOOK_NAME_ON_TRAFFIC_FROM_SERVER priority=1000

3. I tried bytes.Buffer with io.CopyX and a few other methods, but none of them work as expected. They get stuck waiting for an EOF, but PostgreSQL wire protocol doesn't work with EOFs, because it send the message size in the message.
4. I don't want to bind GatewayD to any specific protocol for reading/writing server messages, as it makes it hard to extend and it defeats the purpose of using and extending plugins. That's why I rely on plugins to decode/encode server messages and they need to import the decoder/encoder/parser from the SDK.
5. Setting deadlines for receive and send temporarily helps fix the issue, but it introduces another, that I filed here: Fix bug in handling connection timeouts #199.
6. The last try to use bytes.Buffer worked: Improve client performance #200.

[mostafa]

This is resolved in #343. Pick this up if the issue persists.