connection.close is throwing protocol error

asyncpg version: 0.18.1
PostgreSQL version: 9.5.10
Do you use a PostgreSQL SaaS? If so, which? Can you reproduce
the issue with a local PostgreSQL install?: using google cloud sql pg instance
Python version: 3.6.5
Platform: Ubuntu
Do you use pgbouncer?: No
Did you install asyncpg with pip?: Yes
If you built asyncpg locally, which version of Cython did you use?:
Can the issue be reproduced under both asyncio and
uvloop?: No

I have been facing this error for quite sometime intermittently asyncpg.exceptions._base.InternalClientError: got result for unknown protocol state 0

The entire stack trace:

 File "/repo/server/schema_db.py", line 345, in read
    await connection.close()

  File "/usr/local/lib/python3.6/dist-packages/asyncpg/connection.py", line 1082, in close
    await self._protocol.close(timeout)

  File "asyncpg/protocol/protocol.pyx", line 560, in close

  File "asyncpg/protocol/protocol.pyx", line 814, in asyncpg.protocol.protocol.BaseProtocol._dispatch_result

asyncpg.exceptions._base.InternalClientError: got result for unknown protocol state 0

Can someone explain when this exception is thrown and for what reason?
If i know the reason this exception is thrown, i can figure out the issue myself.
I'm not able to reproduce locally or in my stage environment.

Beginne mit dem fehlschlagenden await connection.close() in schema_db.py:345, untersuche anschließend asyncpg/protocol/protocol.pyx bei close und BaseProtocol._dispatch_result und verwende dabei den gemeldeten Stacktrace und die Versionsdetails. Bestimme die intermittierenden Bedingungen, die mit der Google Cloud SQL PostgreSQL-Instanz den unbekannten Protokollzustand erzeugen; das Issue ist abgeschlossen, wenn die Ursache und ein zuverlässiger Reproduktions- oder Verifizierungsweg festgestellt sind.