feat(detectors): Add filtered URL check to detectors that use urlparse (#97694)

follow up to #97104 - adds the
logic to other detectors

Author: roggenkemper

src/sentry/performance_issues/detectors/consecutive_http_detector.py

@@ -6,7 +6,11 @@
 from sentry.issues.issue_occurrence import IssueEvidence
 from sentry.models.organization import Organization
 from sentry.models.project import Project
-from sentry.performance_issues.detectors.utils import get_max_span_duration, get_total_span_duration
+from sentry.performance_issues.detectors.utils import (
+    get_max_span_duration,
+    get_total_span_duration,
+    has_filtered_url,
+)
 from sentry.utils.event import is_event_from_browser_javascript_sdk
 from sentry.utils.safe import get_path
 
@@ -163,6 +167,9 @@ def _is_eligible_http_span(self, span: Span) -> bool:
         if any([x in description for x in ["_next/static/", "_next/data/", "googleapis.com"]]):
             return False
 
+        if has_filtered_url(self._event, span):
+            return False
+
         return True
 
     def _fingerprint(self) -> str:

src/sentry/performance_issues/detectors/experiments/n_plus_one_api_calls_detector.py

@@ -20,7 +20,7 @@
     parameterize_url,
     parameterize_url_with_result,
 )
-from sentry.performance_issues.detectors.utils import get_total_span_duration
+from sentry.performance_issues.detectors.utils import get_total_span_duration, has_filtered_url
 from sentry.performance_issues.performance_problem import PerformanceProblem
 from sentry.performance_issues.types import Span
 
@@ -128,6 +128,9 @@ def _is_span_eligible(self, span: Span) -> bool:
         if not url:
             return False
 
+        if has_filtered_url(self._event, span):
+            return False
+
         # Once most users update their SDKs to use the latest standard, we
         # won't have to do this, since the URLs will be sent in as `span.data`
         # in a parsed format

src/sentry/performance_issues/detectors/http_overhead_detector.py

@@ -9,6 +9,7 @@
 from sentry.issues.issue_occurrence import IssueEvidence
 from sentry.models.organization import Organization
 from sentry.models.project import Project
+from sentry.performance_issues.detectors.utils import has_filtered_url
 
 from ..base import (
     DetectorType,
@@ -115,13 +116,8 @@ def _is_span_eligible(self, span: Span) -> bool:
             return False
 
         # Check if any spans have filtered URLs
-        event_spans = self._event.get("spans", [])
-        span_index = str(event_spans.index(span) if span in event_spans else -1)
-        meta = self._event.get("_meta", {}).get("spans", {})
-        if span_index in meta:
-            has_filtered_url = meta[span_index].get("data", {}).get("url", {})
-            if has_filtered_url:
-                return False
+        if has_filtered_url(self._event, span):
+            return False
 
         return True
 

src/sentry/performance_issues/detectors/n_plus_one_api_calls_detector.py

@@ -23,7 +23,7 @@
     get_url_from_span,
     parameterize_url,
 )
-from sentry.performance_issues.detectors.utils import get_total_span_duration
+from sentry.performance_issues.detectors.utils import get_total_span_duration, has_filtered_url
 from sentry.performance_issues.performance_problem import PerformanceProblem
 from sentry.performance_issues.types import Span
 
@@ -128,6 +128,10 @@ def _is_span_eligible(self, span: Span) -> bool:
         if not url:
             return False
 
+        # Check if any spans have filtered URLs
+        if has_filtered_url(self._event, span):
+            return False
+
         # Once most users update their SDKs to use the latest standard, we
         # won't have to do this, since the URLs will be sent in as `span.data`
         # in a parsed format

src/sentry/performance_issues/detectors/utils.py

@@ -1,4 +1,5 @@
 import re
+from typing import Any
 
 from sentry.performance_issues.base import get_span_duration
 
@@ -22,3 +23,12 @@ def escape_transaction(transaction: str) -> str:
     transaction = re.sub(r'"', r"\"", transaction)
     transaction = re.sub(r"\*", r"\*", transaction)
     return transaction
+
+
+def has_filtered_url(event: dict[str, Any], span: Span) -> bool:
+    event_spans = event.get("spans", [])
+    span_index = str(event_spans.index(span) if span in event_spans else -1)
+    meta = event.get("_meta", {}).get("spans", {})
+    if span_index in meta and meta[span_index].get("data", {}).get("url", {}):
+        return True
+    return False