Add Log Cache client

[mheath]

This PR adds:

• A Log Cache client
• A Rector implementation of the client
• A test application for testing logging/metrics with Log Cache

The integration tests have been modified to:

• Reuse the test service broker code for pushing the test-log-cache application
• Create Spring Beans for the Log Cache client and test-log-cache metadata
• Log Cache client tests
• Immutables was enabled for the integration tests

[linux-foundation-easycla]

The committers are authorized under a signed CLA.

• ✅ Mike Heath (47a05ba, 855da92)

[twoseat]

Sorry for the possible double-review, I was switching machines!

[mheath]

@twoseat I've made all the changes you suggested. Thanks for being so thorough.

Let me know if you're still seeing intermittent integration test failures. I've run the integration tests 4 or 5 times against my test environment today and haven't seen any Log Cache test failures.

[twoseat]

Picked this up again (sorry for the delay, wrestling with a PR!) and I'm afraid I am still seeing intermittent failures on readCounter. The error I see is:

java.lang.AssertionError: expectation failed (failed running expectation on signal [onNext(ReadResponse{envelopes=EnvelopeBatch{batch=[]}})] with [java.lang.IllegalStateException]: envelope is not present)

I'll keep probing.

[mheath]

@twoseat All the tests are passing now.

[twoseat]

Just a few things and I think this is ready to go! Excellent work, made all the better by it not having to be my work.

[twoseat]

Hi @mheath - I ran the integration tests on this prior to merging and found an issue. While the tests report that they've run fine, the overall log for the run contains multiple references to resource leaks:

io.netty.util.ResourceLeakDetector LEAK: ByteBuf.release() was not called before it's garbage-collected. See https://netty.io/wiki/reference-counted-objects.html for more information.

I've checked a few similar-ish existing tests and don't see the same problem. It's entirely possible that you're just poking an issue in existing code, but if so I haven't been able to track that down. Let me know how you get on recreating it!

Edit: Note that we run with -Dio.netty.leakDetection.level=PARANOID to surface issues like this.

[mheath]

I may be missing something needed to clean things up properly. Let me double check.

[mheath]

This depends on #1084 for the tests to run without any ByteBuf leaks.

[twoseat]

Closed in 5fc354b