Dump stack trace during consumer shutdown

[fpacifici]

We have been seeing a number of cases where, on small workloads, upon consumer restart,
a parallel Kafka Consuemr may hang right after partitions are assigned.
In this scenario the consumer does not pick up any work and, after 30 seconds is kicked
out of the consumer group.

The consumer is then restarted by the liveness probe 5 minutes later.

This PR dumps the stack trace of all threads upon SIGINT if the consumer
is listed in a specific option consumer.dump_stacktrace_on_shutdown.

This way we should be able to see where the consumer gets stuck, thus find a fix.

[codecov]

Codecov Report

❌ Patch coverage is 50.00000% with 4 lines in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
src/sentry/utils/kafka.py	42.85%	4 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           master   #100857      +/-   ##
===========================================
+ Coverage   79.82%    81.14%   +1.31%     
===========================================
  Files        8612      8613       +1     
  Lines      382248    382309      +61     
  Branches    24045     24045              
===========================================
+ Hits       305133    310207    +5074     
+ Misses      76787     71774    -5013     
  Partials      328       328              

[sentry]

Issues attributed to commits in this pull request

This pull request was merged and Sentry observed the following issues:

• ‼️ KafkaException: KafkaError{code=UNKNOWN_MEMBER_ID,... in prod

• ‼️ ChildProcessTerminated: 17 in de

• ‼️ ConsumerError: KafkaError{code=_MAX_POLL_EXCEEDED,... in prod

• ‼️ NoScriptError: Command # 6 (EVALSHA 16dd51e837eac7... in prod

• ‼️ TypeError: '<' not supported between instances of... in prod