Allow ThreadDumpEndpoint to call ThreadMXBean in a native image

[mhalbritter]

At the moment, the ThreadDumpEndpoint is not usable in a native image, as GraalVM doesn't implement thread dumping yet.

I've opened an issue for that: oracle/graal#4714

When that issue is resolved, remove the

if (NativeDetector.inNativeImage()) {
  throw new ThreadDumperUnavailableException("Running in native image");
}

condition.

[wilkinsona]

We're going to remove the condition now so that the fix is available to users as soon as it has been fixed in Graal.

[wilkinsona]

With the condition removed, a client will receive a 500 response:

HTTP/1.1 500 
Connection: close
Content-Type: application/json
Date: Tue, 04 Oct 2022 15:53:34 GMT
Transfer-Encoding: chunked

{
    "error": "Internal Server Error",
    "path": "/actuator/threaddump",
    "status": 500,
    "timestamp": "2022-10-04T15:53:34.629+00:00"
}

And the server will log an UnsupportedFeatureError:

com.oracle.svm.core.jdk.UnsupportedFeatureError: ThreadMXBean methods
	at com.oracle.svm.core.util.VMError.unsupportedFeature(VMError.java:89) ~[na:na]
	at com.oracle.svm.core.jdk.management.SubstrateThreadMXBean.dumpAllThreads(SubstrateThreadMXBean.java:241) ~[na:na]
	at org.springframework.boot.actuate.management.ThreadDumpEndpoint.getFormattedThreadDump(ThreadDumpEndpoint.java:51) ~[actuator-webmvc:3.0.0-SNAPSHOT]
	at org.springframework.boot.actuate.management.ThreadDumpEndpoint.threadDump(ThreadDumpEndpoint.java:42) ~[actuator-webmvc:3.0.0-SNAPSHOT]
	at java.lang.reflect.Method.invoke(Method.java:568) ~[actuator-webmvc:na]
	at org.springframework.util.ReflectionUtils.invokeMethod(ReflectionUtils.java:281) ~[na:na]
	at org.springframework.boot.actuate.endpoint.invoke.reflect.ReflectiveOperationInvoker.invoke(ReflectiveOperationInvoker.java:74) ~[na:na]
	at org.springframework.boot.actuate.endpoint.annotation.AbstractDiscoveredOperation.invoke(AbstractDiscoveredOperation.java:60) ~[actuator-webmvc:3.0.0-SNAPSHOT]
	…

oracle/graal#4714 is tracking removing the limitation from Graal. Until the Graal issue has been fixed, anyone encountering this problem may want to add 👍 reaction to the issue's opening comment. Once it has been fixed, anyone encountering the problem should check that they are using a version of Graal that contains the fix.