grpc: python client often report errors StatusCode.UNAVAILABLE deadline exceeded

What version of gRPC and what language are you using?

vesion: grpcio==1.22.0 language: python

What operating system (Linux, Windows,…) and version?

Linux in docker container

What runtime / compiler are you using (e.g. python version or version of gcc)

python3.6

What did you do?

client and server also use grpc with python, I use a thread to get grpc server’s ip and port from consul. And this is my client code

opts = [ ('grpc.max_message_length', 1024 * 1024 * 1024),
                ('grpc.max_send_message_length', 1024 * 1024 * 1024),
                ('grpc.max_receive_message_length', 1024 * 1024 * 1024),
                 ('grpc.keepalive_time_ms', 10000),
                 ('grpc.keepalive_timeout_ms', 1000),
                 ('grpc.max_connection_idle_ms',1000 * 50)
                 ]
        channel =  grpc.insecure_channel('%s:%d'%(ip, port),opts)
        stub = call_pb2_grpc.RemoteCallStub(channel)
        ct = time.time()
        try:
            result = stub.get_result(call_pb2.Request(method_name=method_name,
                                                      arg=json.dumps(args)
                                                      ), timeout=10,wait_for_ready=True
                                                   ))
        except Exception as e:
            channel.close()

my client and server also use in docker container, and Relations are many-to-many

What did you expect to see?

I hope it can be invoked steadily

What did you see instead?

After running for a period of time，Clients often report errors StatusCode.UNAVAILABLE deadline exceeded

I’ve added service tracking，when I search the record ,I learned that the server did not receive these requests. I run netstat -na|grep ESTABLISHED in client ,See that the connection between client and server is blocked

Anything else we should know about your project / environment?