`zed version` produces error output when spicedb is unavailable

[tstirrat15]

This is likely a regression that we introduced when we introduced retry logic:

We try to connect to the running instance of SpiceDB, and if it isn't available we retry three times. We log on failed retries and this looks weird. We should probably skip the retries and silently fail in this case - otherwise it's just noise.