You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With Vernier installed and configured as the profiler, there is a random segmentation fault encoutered inside Sentry, causing a 500 error user-facing, but no exception ever reported to Sentry. This can cause silent failures where there are no visible signs that the website is flakey other than failed requests.
[9403b-a7c0-4727-b91c-e3518d748965] Started GET "/gems/datadog_api_client/2.23.0/DatadogAPIClient/V2/RUMApplicationCreateRequest.html" for 37.19.207.34 at 2025-02-23 04:27:54 +0000
/app/vendor/bundle/ruby/3.4.0/gems/vernier-1.5.0/lib/vernier/collector.rb:77: [BUG] Segmentation fault at 0x0000000000000000
ruby 3.4.1 (2024-12-25 revision 48d4efcb85) +YJIT +PRISM [x86_64-linux]
-- Ruby level backtrace information ----------------------------------------
/app/vendor/bundle/ruby/3.4.0/gems/puma-6.6.0/lib/puma/thread_pool.rb:167:in 'block in spawn_thread'
/app/vendor/bundle/ruby/3.4.0/gems/puma-6.6.0/lib/puma/server.rb:254:in 'block in run'
/app/vendor/bundle/ruby/3.4.0/gems/puma-6.6.0/lib/puma/server.rb:472:in 'process_client'
/app/vendor/bundle/ruby/3.4.0/gems/puma-6.6.0/lib/puma/request.rb:98:in 'handle_request'
/app/vendor/bundle/ruby/3.4.0/gems/puma-6.6.0/lib/puma/thread_pool.rb:390:in 'with_force_shutdown' /pre>
Reproduction Steps
Unknown.
Expected Behavior
Not to crash.
Actual Behavior
Crashes silently.
Ruby Version
3.4.1
SDK Version
5.22.4
Integration and Its Version
Rails (8.0.1)
Sentry Config
Sentry.initdo |config|
config.dsn=ENV["SENTRY_DSN"]# get breadcrumbs from logsconfig.breadcrumbs_logger=[:active_support_logger,:http_logger]# enable profiling# this is relative to traces_sample_rateconfig.profiles_sample_rate=1.0# Use Vernier for profilingconfig.profiler_class=Sentry::Vernier::Profilerconfig.traces_sampler=lambdado |sampling_context|
# if this is the continuation of a trace, just use that decision (rate controlled by the caller)unlesssampling_context[:parent_sampled].nil?nextsampling_context[:parent_sampled]end0.5endend
The text was updated successfully, but these errors were encountered:
Thanks for reporting this Adam 🙇🏻 I'll be looking into this today, but since this is coming from Vernier I'm afraid I won't be able to fix it quickly. One thing I need to check is why segfaults are not being reported for you by Sentry.
we are simply a library that runs within ruby, segfaults crash ruby entirely so there is no way we could catch those. That would require an agent living parallel to the ruby process.
Issue Description
With Vernier installed and configured as the profiler, there is a random segmentation fault encoutered inside Sentry, causing a 500 error user-facing, but no exception ever reported to Sentry. This can cause silent failures where there are no visible signs that the website is flakey other than failed requests.
I had this manifest on gemdocs.org.
From
Gemfile.lock
Stacktrace
Reproduction Steps
Unknown.
Expected Behavior
Not to crash.
Actual Behavior
Crashes silently.
Ruby Version
3.4.1
SDK Version
5.22.4
Integration and Its Version
Rails (8.0.1)
Sentry Config
The text was updated successfully, but these errors were encountered: