Add micrometer support

[felixbarny]

What does this PR do?

Closes #893

Checklist

• This is an enhancement of existing features, or a new feature in existing plugins
  • I have updated CHANGELOG.asciidoc
  • I have added tests that prove my fix is effective or that my feature works
  • Added an API method or config option? Document in which version this will be introduced
  • I have made corresponding changes to the documentation
• This is a bugfix
  • I have updated CHANGELOG.asciidoc
  • I have added tests that would fail without this fix
• This is a new plugin
  • I have updated CHANGELOG.asciidoc
  • My code follows the style guidelines of this project
  • I have made corresponding changes to the documentation
  • I have added tests that prove my fix is effective or that my feature works
  • New and existing unit tests pass locally with my changes
  • I have updated supported-technologies.asciidoc
  • Added an API method or config option? Document in which version this will be introduced
  • Added an instrumentation plugin? Describe how you made sure that old, non-supported versions are not instrumented by accident.
    We're not instrumenting a particular method in MeterRegistry, just all public methods. The plugin is compiled against 1.0.1 to make sure we don't call methods that have been introduced later. We're only calling public API methods that are guaranteed to be stable in 1.x. Testing against the most recent version doesn't seem necessary. I've tested with the currently most recent version 1.5 manually.
• This is something else
  • I have updated CHANGELOG.asciidoc

[apmmachine]

💚 Build Succeeded

Expand to view the summary

Build stats

• Build Cause: [Pull request #1303 updated]

• Start Time: 2020-08-10T10:08:52.311+0000

• Duration: 49 min 25 sec

Test stats 🧪

Test	Results
Failed	0
Passed	1446
Skipped	11
Total	1457

[felixbarny]

I've also tested this manually with https://github.com/micrometer-metrics/micrometer-samples-spring-boot/

These are the metrics it creates:

{"metricset":{"timestamp":1596121299144000,"samples":{"jvm.gc.memory.allocated":{"value":7987096.0},"tomcat.sessions.created":{"value":0.0},"tomcat.sessions.expired":{"value":0.0},"jvm.gc.max.data.size":{"value":-1.0},"system.cpu.count":{"value":8.0},"process.files.max":{"value":10240.0},"jvm.threads.daemon":{"value":36.0},"process.start.time":{"value":1596121130.193},"tomcat.sessions.active.max":{"value":0.0},"jvm.gc.live.data.size":{"value":0.0},"process.files.open":{"value":110.0},"process.cpu.usage":{"value":0.0014566713814238566},"process.uptime":{"value":169.021},"tomcat.cache.hit":{"value":0.0},"system.load.average.1m":{"value":3.44970703125},"tomcat.cache.access":{"value":0.0},"tomcat.sessions.active.current":{"value":0.0},"system.cpu.usage":{"value":0.1777638190954774},"jvm.threads.live":{"value":38.0},"jvm.classes.loaded":{"value":12398.0},"jvm.classes.unloaded":{"value":0.0},"jvm.threads.peak":{"value":44.0},"jvm.gc.memory.promoted":{"value":0.0},"tomcat.sessions.rejected":{"value":0.0},"tomcat.sessions.alive.max":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"nonheap","id":"CodeHeap 'non-nmethods'"},"samples":{"jvm.memory.committed":{"value":3211264.0},"jvm.memory.max":{"value":7553024.0},"jvm.memory.used":{"value":3158144.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"action":"end of minor GC","cause":"G1 Evacuation Pause"},"samples":{"jvm.gc.pause.count":{"value":2},"jvm.gc.pause.sum.us":{"value":35000.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"heap","id":"G1 Old Gen"},"samples":{"jvm.memory.used":{"value":81345392.0},"jvm.memory.committed":{"value":137363456.0},"jvm.memory.max":{"value":4294967296.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"nonheap","id":"Metaspace"},"samples":{"jvm.memory.max":{"value":-1.0},"jvm.memory.committed":{"value":65265664.0},"jvm.memory.used":{"value":61225784.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"level":"trace"},"samples":{"logback.events":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"level":"info"},"samples":{"logback.events":{"value":43.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"exception":"None","method":"GET","phi":"0.95","status":"200","uri":"/api/people"},"samples":{"http.server.requests.percentile":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"heap","id":"G1 Eden Space"},"samples":{"jvm.memory.committed":{"value":125829120.0},"jvm.memory.used":{"value":52428800.0},"jvm.memory.max":{"value":-1.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"level":"warn"},"samples":{"logback.events":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"id":"direct"},"samples":{"jvm.buffer.memory.used":{"value":1176295.0},"jvm.buffer.total.capacity":{"value":1176295.0},"jvm.buffer.count":{"value":3.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"level":"debug"},"samples":{"logback.events":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"name":"http-nio-8080"},"samples":{"tomcat.global.request.max":{"value":0.43},"tomcat.global.error":{"value":0.0},"tomcat.global.sent":{"value":15.0},"tomcat.threads.config.max":{"value":200.0},"tomcat.global.received":{"value":0.0},"tomcat.threads.busy":{"value":0.0},"tomcat.threads.current":{"value":10.0},"tomcat.global.request.count":{"value":1},"tomcat.global.request.sum.us":{"value":430000.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"exception":"None","method":"GET","phi":"0.999","status":"200","uri":"/api/people"},"samples":{"http.server.requests.percentile":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"nonheap","id":"CodeHeap 'non-profiled nmethods'"},"samples":{"jvm.memory.max":{"value":244105216.0},"jvm.memory.committed":{"value":14155776.0},"jvm.memory.used":{"value":14121216.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"level":"error"},"samples":{"logback.events":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"exception":"None","method":"GET","status":"200","uri":"/api/people"},"samples":{"http.server.requests.count":{"value":1},"http.server.requests.sum.us":{"value":293042.45}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"id":"mapped"},"samples":{"jvm.buffer.count":{"value":0.0},"jvm.buffer.total.capacity":{"value":0.0},"jvm.buffer.memory.used":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"action":"end of minor GC","cause":"Metadata GC Threshold"},"samples":{"jvm.gc.pause.count":{"value":1},"jvm.gc.pause.sum.us":{"value":17000.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"heap","id":"G1 Survivor Space"},"samples":{"jvm.memory.committed":{"value":5242880.0},"jvm.memory.max":{"value":-1.0},"jvm.memory.used":{"value":5242880.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"name":"default"},"samples":{"tomcat.servlet.request.count":{"value":0},"tomcat.servlet.request.sum.us":{"value":0.0},"tomcat.servlet.error":{"value":0.0},"tomcat.servlet.request.max":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"exception":"None","method":"GET","phi":"0.5","status":"200","uri":"/api/people"},"samples":{"http.server.requests.percentile":{"value":0.0}}}}
{"metricset":{"timestamp":1596121299144000,"tags":{"area":"nonheap","id":"Compressed Class Space"},"samples":{"jvm.memory.max":{"value":1073741824.0},"jvm.memory.used":{"value":8402928.0},"jvm.memory.committed":{"value":9175040.0}}}}

One issue is that the metrics starting with http are not indexed because APM Server applies common fields to all indices:
https://github.com/elastic/apm-server/blob/57830070b4bbc1bb487eefd96bb79fb9bc203901/_meta/fields.common.yml#L72-L74

This is intentional. Quoting @graphaelli:

so yes, the metric index getting the same mapping as the others is intentional, splitting indices by type is optional though the default

[felixbarny]

As docs go, we might want to add a section on how to visualize a timer and a gauge with Lens and/or TSVB. I suspect visualizing a timer may be problematic with Lens as we have to calculate a weighted average based on the .count and the .sum.us part of the timer.

[eyalkoren]

Started looking into it, but there is a consistent failure in integration tests at its current state

[felixbarny]

Thx for the notice. Should be fixed with the latest commit. Let's see...

[eyalkoren]

Beautiful!
A few minor comments.
Main one - why do we serialize this differently (to a buffer first and then to the actual connection)?

[eyalkoren]

🎉

[felixbarny]

Histogram support will be added later once elastic/apm-server#3195 is resolved