Improve Timescale instance for "Too many clients already"

[tomjohn1028]

The following error was showing up in some Hasura logs

{"type":"http-log","timestamp":"2021-09-09T01:34:32.766+0000","level":"error","detail":{"operation":{"user_vars":{"x-hasura-role":"admin"},"error":{"internal":"FATAL:  sorry, too many clients already
","path":"$","error":"connection error","code":"postgres-error"},"request_id":"d8de1115-fa19-41ff-b5c9-149a78b9b095","response_size":118,"query":{"variables":{"object":{"order_id":"46891623435369680254532","client_id":"9254101784418634862","serum_market_address":"qDvaLApMYkbp9NTgYJaNy7CDsPhXMbDjEr3g3oqEMaw","account":"Abe39jLaH5BSmLryRbAGu1wMfCgPkXnvkQV7JZDuvDxT","slot":95649683,"account_slot":2,"type":"done","side":"sell","timestamp":"2021-09-09T01:33:51.437Z"}},"query":"
      mutation ($object: serum_events_insert_input!) {
        insert_serum_events_one(object: $object) {
            serum_market_address
        }
      }
      
        "}},"http_info":{"status":200,"http_version":"HTTP/1.1","url":"/v1/graphql","ip":"3.70.45.116","method":"POST","content_encoding":null}}}

This seems to coincide with when a new Hasura task is being spun up and the old one is being shut down (see image below for hasura timing).

And the following for Timescale's last error logs.

This makes sense as the Hasura connection pool would surpass the postgres max_connections if two were running at the same time and they used the Timescale tuning to determine max connections.

[tomjohn1028]

More insight to Hasura tuning and connections https://hgiasac.github.io/posts/2020-05-15-Tuning-Hasura-at-Scale.html#hasura-graphql-pg-connections