A ugly design

[MegaByte875]

If an RPC service endpoint is not started, the calling service will crash, wouldn't this couple them together? In a production environment, if service A depends on services B and C, and at some point, service B or C fails and becomes unreachable for a period of time, should service A also detect this and exit?

[kevwan]

It depends on your logic, this is fail-fast strategy.

But still you can set NonBlock: true to change the behavior.

BTW: this is by design, and I don't think it's a ugly design.

[zhqqqy]

or add an exponential backoff retry mechanism?

It depends on your logic, this is fail-fast strategy.

But still you can set NonBlock: true to change the behavior.

BTW: this is by design, and I don't think it's a ugly design.