-
Notifications
You must be signed in to change notification settings - Fork 373
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Upgrade profiler to use libdatadog v2.0.0 #2599
Conversation
…tring This is a newly-added feature to both the profiling backend, as well as libdatadog. Instead of sending the local root span ids as strings (that take up space in the string table, etc), we can now send them as numbers. We already did a similar migration for span ids in #2476, but at the time we did not change the local root span ids because the libdatadog `ddog_prof_Profile_set_endpoint` API could not properly handle the numeric ids (this has been changed in DataDog/libdatadog#80 ). As far as testing goes, I could have chosen to abstract away this change from our specs BUT I chose to make it explicit and changed the specs to match. I like having visibility in the specs of the values being encoded and passed around as numbers and not strings.
This will allow us to have early warning for when we're using the API incorrectly. See also DataDog/libdatadog#87 .
One of the drop methods had to be commented out due to a libdatadog bug, I'll re-enable it once libdatadog gets fixed.
if (push_result.tag == DDOG_VEC_TAG_PUSH_RESULT_ERR) { | ||
VALUE err_details = ruby_string_from_vec_u8(push_result.err); | ||
ddog_Vec_Tag_PushResult_drop(push_result); | ||
|
||
// libdatadog validates tags and may catch invalid tags that ddtrace didn't actually catch. | ||
// We warn users about such tags, and then just ignore them. | ||
safely_log_failure_to_process_tag(tags, err_details); | ||
} else { | ||
ddog_Vec_Tag_PushResult_drop(push_result); | ||
safely_log_failure_to_process_tag(tags, get_error_details_and_drop(&push_result.err)); | ||
} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The else branch is not actually needed because there's nothing to drop when the ddog_Vec_Tag_push
operation was successful.
VALUE encoded_pprof = ruby_string_from_prof_vec_u8(serialized_profile.ok.buffer); | ||
VALUE encoded_pprof = ruby_string_from_vec_u8(serialized_profile.ok.buffer); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This was a workaround for a header issue in 1.0.1, and it's been fixed.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks good to me. I did more than a glance, but not a particularly thorough review. If you would like more scrutiny, throw some time on calendar to go over it together!
Sounds good, thanks for the review! |
What does this PR do?:
This PR updates the profiler to use the new libdatadog v2.0.0 APIs.
In particular:
Motivation:
Enable the Ruby profiler to use the latest libdatadog release.
Additional Notes:
A few of the changes were committed incrementally, so it may be easier to review commit-by-commit.
How to test the change?:
The changes are covered by our existing test suite. In fact, this branch helped pinpoint a libdatadog bug that was fixed before release.