Consider reusing HazelcastHealthIndicator for more consistent results

[snicoll]

Rather than having a custom HazelcastJetHealthIndicator, please consider auto-configuring a standard org.springframework.boot.actuate.hazelcast.HazelcastHealthIndicator since all it does is checking the state of the embdded HazelcastInstance anyway.

[gurbuzali]

org.springframework.boot.actuate.hazelcast.HazelcastHealthIndicator uses com.hazelcast.core.Endpoint which has moved to com.hazelcast.cluster.Endpoint in 4.0. This prevents us to reuse HazelcastHealthIndicator. But if we change it to use the LifecycleService.isRunning instead of executing a transaction, we can re-visit this issue.

[snicoll]

Ah good point, I haven't started to look at Hazelcast v4 compatibility yet. Perhaps you could take the Spring Boot version here and port it to the new API rather than creating a completely different one?

We can coordinate once Spring Boot upgrades to Hazelcast v4.