This is maybe a duplicate of #19161, unfortunately that ticket did not mention where/how to "increase the wait time" (DEFAULT_POLL_INTERVAL).

We're getting errors like:

"Connect failed: {"created":"@1582198742.973444006","description":"Failed to connect to remote host: FD Shutdown","file":"third_party/grpc/src/core/lib/iomgr/lockfree_event.cc","file_line":195,"os_error":"Timeout occurred","referenced_errors":[{"created":"@1582198742.973420053","description":"connect() timed out","file":"third_party/grpc/src/core/lib/iomgr/tcp_client_posix.cc","file_line":112}],"target_address":"ipv4:10.100.214.244:12346"}
"

where the client is retrying a dozen times and then it fails.

In our case we have a grpc server that gets the initial call from a remote client. The case were the error happens is a nested call, where the server invokes itself. This works when we run the server standalone, but fails when it is running in a container.