Right now, ServiceQuery#using(), #resolveAll() and #resolveOne() are attempted once and simply fail if the requested services are not available. There should be a way to configure each ServiceQuery to reattempt resolution until services matching some arbitrary criteria become available. A reasonable default criteria could be until the service becomes available. I guess there should be ways to limit the number of attempts, or the duration/deadline after which reattempting should stop.
When we're at it, there should probably be a way to specify that any caches should be circumvented.
Right now, ServiceQuery#using(), #resolveAll() and #resolveOne() are attempted once and simply fail if the requested services are not available. There should be a way to configure each ServiceQuery to reattempt resolution until services matching some arbitrary criteria become available. A reasonable default criteria could be until the service becomes available. I guess there should be ways to limit the number of attempts, or the duration/deadline after which reattempting should stop.
When we're at it, there should probably be a way to specify that any caches should be circumvented.