Mark ddtrace threads as fork-safe #3279
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
**What does this PR do?** This PR applies to all relevant ddtrace threads a pattern pioneered by rails: set the `:fork_safe` thread variable to indicate that it's OK if ddtrace threads are started before a Ruby app calls fork. **Motivation:** This PR fixes the issue reported in #3203. The puma webserver in clustering mode has a check where it will warn when it detects that threads were started before the webserver has fork'd the child processes, see: https://github.com/puma/puma/blob/32e011ab9e029c757823efb068358ed255fb7ef4/lib/puma/cluster.rb#L353-L359 This is not a problem for ddtrace (as discussed in #3203), but the warning is annoying. While looking into the issue again today, I spotted that actually there's a way to tell puma that a thread is fine and is it's not a problem if it's started before the webserver has fork'd: puma checks if the thread has a `:fork_safe` thread-variable set to `true`. **Additional Notes:** We have a bunch of places in ddtrace where we create threads... I briefly played with the idea of creating a superclass of all ddtrace Threads, but decided to not get into a bigger refactoring for such a small issue. Maybe next time...? **How to test the change?** Test coverage included. Furthermore, you can try running a Ruby webapp in puma before/after the change to confirm the warning is gone. Fixes #3203
github-actions
bot
added
core
Closed
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## master #3279 +/- ##
==========================================
- Coverage 98.23% 98.22% -0.01%
==========================================
Files 1253 1253
Lines 72393 72418 +25
Branches 3393 3397 +4
==========================================
+ Hits 71112 71133 +21
- Misses 1281 1285 +4 ☔ View full report in Codecov by Sentry. |
marcotc
approved these changes
Nov 23, 2023
Done in e430e78 :) |
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
This PR applies to all relevant ddtrace threads a pattern pioneered by rails: set the
:fork_safethread variable to indicate that it's OK if ddtrace threads are started before a Ruby app calls fork.Motivation:
This PR fixes the issue reported in #3203. The puma webserver in clustering mode has a check where it will warn when it detects that threads were started before the webserver has fork'd the child processes, see:
https://github.com/puma/puma/blob/32e011ab9e029c757823efb068358ed255fb7ef4/lib/puma/cluster.rb#L353-L359
This is not a problem for ddtrace (as discussed in #3203), but the warning is annoying.
While looking into the issue again today, I spotted that actually there's a way to tell puma that a thread is fine and is it's not a problem if it's started before the webserver has fork'd: puma checks if the thread has a
:fork_safethread-variable set totrue.Additional Notes:
We have a bunch of places in ddtrace where we create threads... I briefly played with the idea of creating a superclass of all ddtrace Threads, but decided to not get into a bigger refactoring for such a small issue. Maybe next time...?
How to test the change?
Test coverage included. Furthermore, you can try running a Ruby webapp in puma before/after the change to confirm the warning is gone.
For Datadog employees:
credentials of any kind, I've requested a review from
@DataDog/security-design-and-guidance.Fixes #3203