fix(profiling): detect and correct non-string Location fields [backport 1.20] (#7010)

Backport 262a336 from #6675 to 1.20.

Fixes #6652 by skipping objects that are not strings. This appears to be
behavior specific to python 3.11, similar to how we have to fence
frame/code objects.

Still not 100% sure what causes this to happen, as we don't have a repro
case. The sanitization procedure _could_ be special-cased to 3.11 or
later, as we have no evidence of this defect in earlier pythons.

## Checklist

- [X] Change(s) are motivated and described in the PR description.
- [X] Testing strategy is described if automated tests are not included
in the PR.
- [X] Risk is outlined (performance impact, potential for breakage,
maintainability, etc).
- [X] Change is maintainable (easy to change, telemetry, documentation).
- [X] [Library release note
guidelines](https://ddtrace.readthedocs.io/en/stable/releasenotes.html)
are followed. If no release note is required, add label
`changelog/no-changelog`.
- [X] Documentation is included (in-code, generated user docs, [public
corp docs](https://github.com/DataDog/documentation/)).
- [X] Backport labels are set (if
[applicable](https://ddtrace.readthedocs.io/en/latest/contributing.html#backporting))

## Reviewer Checklist

- [x] Title is accurate.
- [x] No unnecessary changes are introduced.
- [x] Description motivates each change.
- [x] Avoids breaking
[API](https://ddtrace.readthedocs.io/en/stable/versioning.html#interfaces)
changes unless absolutely necessary.
- [x] Testing strategy adequately addresses listed risk(s).
- [x] Change is maintainable (easy to change, telemetry, documentation).
- [x] Release note makes sense to a user of the library.
- [x] Reviewer has explicitly acknowledged and discussed the performance
implications of this PR as reported in the benchmarks PR comment.
- [x] Backport labels are set in a manner that is consistent with the
[release branch maintenance
policy](https://ddtrace.readthedocs.io/en/latest/contributing.html#backporting)

Co-authored-by: David Sanchez <[email protected]>

Author: github-actions[bot]

ddtrace/internal/datadog/profiling/_ddup.pyx

@@ -8,6 +8,8 @@ from ddtrace.internal.compat import ensure_binary
 from ddtrace.internal.constants import DEFAULT_SERVICE_NAME
 from ddtrace.span import Span
 
+from .utils import sanitize_string
+
 
 IF UNAME_SYSNAME == "Linux":
     cdef extern from "exporter.hpp":
@@ -115,8 +117,8 @@ IF UNAME_SYSNAME == "Linux":
         ddup_push_lock_name(ensure_binary(lock_name))
 
     def push_frame(name: str, filename: str, address: int, line: int) -> None:
-        name = name if name is not None else ""
-        filename = filename if filename is not None else ""
+        name = sanitize_string(name)
+        filename = sanitize_string(filename)
         ddup_push_frame(ensure_binary(name), ensure_binary(filename), address, line)
 
     def push_threadinfo(thread_id: int, thread_native_id: int, thread_name: Optional[str]) -> None:

ddtrace/internal/datadog/profiling/ddup.py

@@ -1,3 +1,6 @@
+from .utils import sanitize_string  # noqa: F401
+
+
 try:
     from ._ddup import *  # noqa: F403, F401
 except ImportError:

ddtrace/internal/datadog/profiling/utils.py

@@ -0,0 +1,31 @@
+from sys import version_info
+from typing import Any
+
+from ddtrace.internal.logger import get_logger
+
+
+LOG = get_logger(__name__)
+
+
+# 3.11 and above
+def _sanitize_string_check(value):
+    # type: (Any) -> str
+    if isinstance(value, str):
+        return value
+    elif value is None:
+        return ""
+    try:
+        return value.decode("utf-8", "ignore")
+    except Exception:
+        LOG.warning("Got object of type '%s' instead of str during profile serialization", type(value).__name__)
+        return "[invalid string]%s" % type(value).__name__
+
+
+# 3.10 and below (the noop version)
+def _sanitize_string_identity(value):
+    # type: (Any) -> str
+    return value or ""
+
+
+# Assign based on version
+sanitize_string = _sanitize_string_check if version_info[:2] > (3, 10) else _sanitize_string_identity

ddtrace/profiling/exporter/pprof.pyx

@@ -12,6 +12,7 @@ from ddtrace import ext
 from ddtrace.internal import packages
 from ddtrace.internal._encoding import ListStringTable as _StringTable
 from ddtrace.internal.compat import ensure_str
+from ddtrace.internal.datadog.profiling.ddup import sanitize_string
 from ddtrace.internal.utils import config
 from ddtrace.profiling import event
 from ddtrace.profiling import exporter
@@ -190,9 +191,14 @@ class _PprofConverter(object):
 
     def _to_Function(
         self,
-        filename: str,
-        funcname: str,
-    ) -> pprof_FunctionType:
+        filename,  # type: str
+        funcname,  # type: str
+    ):
+        # type: (...) -> pprof_FunctionType
+        # filename/funcname are "guaranteed" to be str, but on 3.11 and later
+        # they may (erroneously?) be bytes.  Try to fix this.
+        filename = sanitize_string(filename)
+        funcname = sanitize_string(funcname)
         try:
             return self._functions[(filename, funcname)]
         except KeyError:
@@ -206,10 +212,15 @@ class _PprofConverter(object):
 
     def _to_Location(
         self,
-        filename: str,
-        lineno: int,
-        funcname: str,
-    ) -> pprof_LocationType:
+        filename,  # type: str
+        lineno,  # type: int
+        funcname,  # type: str
+    ):
+        # type: (...) -> pprof_LocationType
+        # filename/funcname are "guaranteed" to be str, but on 3.11 and later
+        # they may (erroneously?) be bytes.  Try to fix this.
+        filename = sanitize_string(filename)
+        funcname = sanitize_string(funcname)
         try:
             return self._locations[(filename, lineno, funcname)]
         except KeyError:

releasenotes/notes/fix-profiler-frame-strings-24a10dad43f25bc7.yaml

@@ -0,0 +1,5 @@
+---
+fixes:
+  - |
+    profiling: Fixed an issue with data encoding where non-string objects
+    might be interned on Python 3.11