feat(data-collection): Request data collection by sl0thentr0py · Pull Request #3030 · getsentry/sentry-ruby

sl0thentr0py · 2026-07-21T14:35:18Z

Add PII_HEADER_SNIPPETS to backfill
Use data_collection.http_headers.request for ip address collection (both env and request)
- Sensitive headers like auth and cookies are now ALWAYS filtered unlike before
Use data_collection.cookies for cookie collection (fix Hash signature since it always was one)
Use data_collection.url_query_params for query string and parse query into Hash and filter
Use data_collection.http_bodies for body
- JSON is also parsed and filtered now, form data is also filtered for sensitive keys
Deprecate rack_env_whitelist, will be removed in major

Issues

Resolves: feat(pii): Wire core request interface to data_collection #3003
Resolves: RUBY-191

linear-code · 2026-07-21T15:23:30Z

cursor

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 831da33. Configure here.}

Co-Authored-By: OpenAI <noreply@example.com>

solnic · 2026-07-30T11:30:53Z


+    # Filters key-value data using the default sensitive denylist.
+    def self.filter(values)
+      KeyValueCollection.new(mode: :deny_list, terms: nil).filter(values)


I'd memoize this, no need to keep creating the same instance all over again.

def self.filter(values) default_filter.filter(values) end def self.default_filter @default_filter ||= KeyValueColletion.new(mode: :deny_list, terms: nil) end

solnic · 2026-07-30T11:32:18Z

        collection.terms = ["USER"]

-        expect(collection.filter("user_id" => "1")).to eq("user_id" => "[Filtered]")
+        expect(collection.filter({ "user_id" => "1" })).to eq("user_id" => "[Filtered]")


What was wrong with the implicit form? 🤔

jruby bug with kwargs

solnic · 2026-07-30T11:34:25Z

          headers: { 'Host' => 'localhost', "X-Forwarded-For" => "1.1.1.1, 2.2.2.2", "X-Request-Id" => "abcd-1234-abcd-1234" },
          method: 'POST',
-          query_string: 'biz=baz',
+          query_string: { 'biz' => 'baz' },


Wait, this is no longer a string - is this correct?

yes, I will test through with the product, but since relay would've parsed it either way, behavior should remain the same eventually (except for before_send which we will call out in the changelog)

cursor Bot reviewed Jul 21, 2026

View reviewed changes

Comment thread sentry-ruby/lib/sentry/interfaces/request.rb

sl0thentr0py force-pushed the neel/data-collection/request branch 4 times, most recently from 84ceec2 to e36c648 Compare July 21, 2026 15:21

sl0thentr0py force-pushed the neel/data-collection/request branch 2 times, most recently from ceb85b7 to 6a366a2 Compare July 22, 2026 11:39

cursor Bot reviewed Jul 22, 2026

View reviewed changes

Comment thread sentry-ruby/lib/sentry/interfaces/request.rb Outdated