Test indy against full test matrix

[trask]

I'm getting failures when running indy against openj9 locally, not sure if it's a real issue or not, but realized we're missing some test coverage.

related to #13031

cc @JonasKunz @SylvainJuge

[JonasKunz]

At first sight looks like the fix for recursive invokedynamic-bootstrapping implemented in #12242 seems to work only for hotspot and not openJ9.

        at java.lang.invoke.MutableCallSite.<init>(MutableCallSite.java:77)
	at io.opentelemetry.javaagent.tooling.instrumentation.indy.IndyBootstrap.bootstrapAdvice(IndyBootstrap.java:183)
	at io.opentelemetry.javaagent.tooling.instrumentation.indy.IndyBootstrap.internalBootstrap(IndyBootstrap.java:135)
	at io.opentelemetry.javaagent.tooling.instrumentation.indy.IndyBootstrap.bootstrap(IndyBootstrap.java:115)
	at io.opentelemetry.javaagent.bootstrap.IndyBootstrapDispatcher.bootstrap(IndyBootstrapDispatcher.java:43)
	at java.lang.invoke.MethodHandleResolver.invokeBsm(MethodHandleResolver.java:117)
	at java.lang.invoke.MethodHandleResolver.resolveInvokeDynamic(MethodHandleResolver.java:165)
	at com.ibm.oti.vm.BootstrapClassLoader.loadClass(BootstrapClassLoader.java:79)
	at java.lang.invoke.MutableCallSite.<init>(MutableCallSite.java:77)
	at io.opentelemetry.javaagent.tooling.instrumentation.indy.IndyBootstrap.bootstrapAdvice(IndyBootstrap.java:183)
	at io.opentelemetry.javaagent.tooling.instrumentation.indy.IndyBootstrap.internalBootstrap(IndyBootstrap.java:135)
	at io.opentelemetry.javaagent.tooling.instrumentation.indy.IndyBootstrap.bootstrap(IndyBootstrap.java:115)
	at io.opentelemetry.javaagent.bootstrap.IndyBootstrapDispatcher.bootstrap(IndyBootstrapDispatcher.java:43)
	at java.lang.invoke.MethodHandleResolver.invokeBsm(MethodHandleResolver.java:117)
	at java.lang.invoke.MethodHandleResolver.resolveInvokeDynamic(MethodHandleResolver.java:165)
	at com.ibm.oti.vm.BootstrapClassLoader.loadClass(BootstrapClassLoader.java:79)
	at java.lang.invoke.MutableCallSite.<init>(MutableCallSite.java:77)

The problem seems to be that on the code path that should prevent the recursion we trigger a new recursion:
The construction of MutableCallSite causes classloading, which again involves our instrumentation.

Best case is we should get away with just creating a MutableCallSite in the static initializer of IndyBootstrap to eagerly trigger the invovled classloading to remove it from the recursion code path.
I'll give it a try.