properly set thread name in thread context

[colinsurprenant]

@guyboertje recently ran through me a bug he discovered in the workers thread naming which was introduced in #9973. I decided to submit this PR because I'd like to get this fix in before I go ahead and rebase and make progress with #9991.

The problem is that the call to Util.set_thread_name was made outside the newly created thread context and the Util.set_thread_name method set the passed name to the current thread context.

The problem is that none of the worker threads were given a name and the pipeline thread name was set to the last worker thread name.

This problem was present on both Ruby and Java pipelines and for both cases using the hot thread API to gather the thread names this is the before/after:

curl -XGET 'localhost:9600/_node/hot_threads?human=true&ignore_idle_threads=false&threads=100' | grep "thread name:"

Before

Note the single [main]>worker7 thread (which is in fact the pipeline-manager thread) and the many Ruby-0-Thread-X threads:

118.77 % of cpu usage, state: timed_waiting, thread name: 'LogStash::Runner', thread id: 1
4.79 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-8', thread id: 24
2.07 % of cpu usage, state: timed_waiting, thread name: 'Api Webserver', thread id: 32
0.98 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-1', thread id: 16
0.88 % of cpu usage, state: timed_waiting, thread name: 'pool-3-thread-1', thread id: 18
0.77 % of cpu usage, state: timed_waiting, thread name: '[main]>worker7', thread id: 19
0.42 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-7', thread id: 23
0.42 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-10', thread id: 26
0.41 % of cpu usage, state: timed_waiting, thread name: 'pool-4-thread-1', thread id: 15
0.37 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-6', thread id: 22
0.35 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-11', thread id: 27
0.35 % of cpu usage, state: waiting, thread name: 'Finalizer', thread id: 3
0.34 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-4', thread id: 20
0.32 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-9', thread id: 25
0.31 % of cpu usage, state: waiting, thread name: 'Reference Handler', thread id: 2
0.31 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-5', thread id: 21
0.31 % of cpu usage, state: runnable, thread name: 'Ruby-0-Thread-17', thread id: 36
0.21 % of cpu usage, state: runnable, thread name: '[main]<stdin', thread id: 30
0.14 % of cpu usage, state: runnable, thread name: 'Ruby-0-Thread-14', thread id: 33
0.14 % of cpu usage, state: timed_waiting, thread name: 'pool-3-thread-2', thread id: 28
0.09 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-15', thread id: 34
0.07 % of cpu usage, state: timed_waiting, thread name: 'pool-3-thread-3', thread id: 29
0.05 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-16', thread id: 35
0.0 % of cpu usage, state: timed_waiting, thread name: 'logstash-pipeline-flush', thread id: 31
0.0 % of cpu usage, state: runnable, thread name: 'Signal Dispatcher', thread id: 5

After

Note all the proper [main]>workerX threads and the [main]-pipeline-manager thread:

92.38 % of cpu usage, state: timed_waiting, thread name: 'LogStash::Runner', thread id: 1
3.84 % of cpu usage, state: timed_waiting, thread name: '[main]>worker1', thread id: 20
1.64 % of cpu usage, state: timed_waiting, thread name: 'Api Webserver', thread id: 30
0.79 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-1', thread id: 16
0.67 % of cpu usage, state: timed_waiting, thread name: 'pool-3-thread-1', thread id: 29
0.56 % of cpu usage, state: timed_waiting, thread name: '[main]-pipeline-manager', thread id: 18
0.41 % of cpu usage, state: timed_waiting, thread name: '[main]>worker3', thread id: 22
0.37 % of cpu usage, state: timed_waiting, thread name: '[main]>worker4', thread id: 23
0.35 % of cpu usage, state: timed_waiting, thread name: '[main]>worker6', thread id: 25
0.34 % of cpu usage, state: timed_waiting, thread name: '[main]>worker2', thread id: 21
0.34 % of cpu usage, state: timed_waiting, thread name: 'pool-4-thread-1', thread id: 15
0.31 % of cpu usage, state: timed_waiting, thread name: '[main]>worker7', thread id: 26
0.31 % of cpu usage, state: timed_waiting, thread name: '[main]>worker0', thread id: 19
0.3 % of cpu usage, state: timed_waiting, thread name: '[main]>worker5', thread id: 24
0.28 % of cpu usage, state: waiting, thread name: 'Finalizer', thread id: 3
0.26 % of cpu usage, state: waiting, thread name: 'Reference Handler', thread id: 2
0.24 % of cpu usage, state: runnable, thread name: 'Ruby-0-Thread-17', thread id: 36
0.16 % of cpu usage, state: runnable, thread name: '[main]<stdin', thread id: 27
0.11 % of cpu usage, state: runnable, thread name: 'Ruby-0-Thread-14', thread id: 33
0.11 % of cpu usage, state: timed_waiting, thread name: 'pool-3-thread-3', thread id: 32
0.06 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-15', thread id: 34
0.03 % of cpu usage, state: timed_waiting, thread name: 'Ruby-0-Thread-16', thread id: 35
0.03 % of cpu usage, state: timed_waiting, thread name: 'pool-3-thread-2', thread id: 31
0.0 % of cpu usage, state: timed_waiting, thread name: 'logstash-pipeline-flush', thread id: 28
0.0 % of cpu usage, state: runnable, thread name: 'Signal Dispatcher', thread id: 5

[danhermann]

LGTM pending green CI

[colinsurprenant]

Great, thanks @danhermann - looking at backports.

[colinsurprenant]

backported in 6.4, 6.5, 6.6 and 6.x