datadog tracer cause high cpu issue.

[ihangji]

datadog-tracer-java encountered a problem, and then would keep looping inside that problem. This leads to high cpu usage.

[1] Busy(99.9%) thread(56/0x38) stack of java process(7) under user(jetty): "dd-task-scheduler" #42 daemon prio=5 os_prio=0 tid=0x00007f0c6315d800 nid=0x38 runnable [0x00007f0ba2415000] java.lang.Thread.State: RUNNABLE at datadog.trace.common.metrics.ConflatingMetricsAggregator.report(ConflatingMetricsAggregator.java:131) at datadog.trace.common.metrics.ConflatingMetricsAggregator$ReportTask.run(ConflatingMetricsAggregator.java:245) at datadog.trace.common.metrics.ConflatingMetricsAggregator$ReportTask.run(ConflatingMetricsAggregator.java:240) at datadog.trace.util.AgentTaskScheduler$PeriodicTask.run(AgentTaskScheduler.java:268) at datadog.trace.util.AgentTaskScheduler$Worker.run(AgentTaskScheduler.java:223) at java.lang.Thread.run(Thread.java:748)

[mcculls]

Hi @ihangji - which version of the Java tracer are you using?

Based on the stack trace this looks like a known issue where metrics aggregation was incorrectly enabled in the tracer when the agent version didn't support it, leading to high CPU usage.

The nearest version with the fix that doesn't have other known issues is 0.78.3, but I would recommend trying 0.88.0 if you can because it contains our latest fixes and several integration improvements.