feat(profiling): Add profiler options to init (#1947)

This adds the `profiles_sample_rate`, `profiles_sampler` and `profiler_mode`
options to the top level of the init call. The `_experiment` options will still
be available temporarily but is deprecated and will be removed in the future.

Author: Zylphrex

sentry_sdk/_types.py

@@ -85,3 +85,5 @@
 
     FractionUnit = Literal["ratio", "percent"]
     MeasurementUnit = Union[DurationUnit, InformationUnit, FractionUnit, str]
+
+    ProfilerMode = Literal["sleep", "thread", "gevent", "unknown"]

sentry_sdk/client.py

@@ -28,7 +28,7 @@
 from sentry_sdk.utils import ContextVar
 from sentry_sdk.sessions import SessionFlusher
 from sentry_sdk.envelope import Envelope
-from sentry_sdk.profiler import setup_profiler
+from sentry_sdk.profiler import has_profiling_enabled, setup_profiler
 
 from sentry_sdk._types import TYPE_CHECKING
 
@@ -174,8 +174,7 @@ def _capture_envelope(envelope):
         finally:
             _client_init_debug.set(old_debug)
 
-        profiles_sample_rate = self.options["_experiments"].get("profiles_sample_rate")
-        if profiles_sample_rate is not None and profiles_sample_rate > 0:
+        if has_profiling_enabled(self.options):
             try:
                 setup_profiler(self.options)
             except ValueError as e:

sentry_sdk/consts.py

@@ -19,6 +19,7 @@
         BreadcrumbProcessor,
         Event,
         EventProcessor,
+        ProfilerMode,
         TracesSampler,
         TransactionProcessor,
     )
@@ -33,8 +34,9 @@
             "max_spans": Optional[int],
             "record_sql_params": Optional[bool],
             "smart_transaction_trimming": Optional[bool],
+            # TODO: Remvoe these 2 profiling related experiments
             "profiles_sample_rate": Optional[float],
-            "profiler_mode": Optional[str],
+            "profiler_mode": Optional[ProfilerMode],
         },
         total=False,
     )
@@ -115,6 +117,9 @@ def __init__(
         propagate_traces=True,  # type: bool
         traces_sample_rate=None,  # type: Optional[float]
         traces_sampler=None,  # type: Optional[TracesSampler]
+        profiles_sample_rate=None,  # type: Optional[float]
+        profiles_sampler=None,  # type: Optional[TracesSampler]
+        profiler_mode=None,  # type: Optional[ProfilerMode]
         auto_enabling_integrations=True,  # type: bool
         auto_session_tracking=True,  # type: bool
         send_client_reports=True,  # type: bool

sentry_sdk/profiler.py

@@ -27,6 +27,7 @@
 from sentry_sdk._types import TYPE_CHECKING
 from sentry_sdk.utils import (
     filename_for_module,
+    is_valid_sample_rate,
     logger,
     nanosecond_time,
     set_in_app_in_frames,
@@ -46,7 +47,7 @@
     from typing_extensions import TypedDict
 
     import sentry_sdk.tracing
-    from sentry_sdk._types import SamplingContext
+    from sentry_sdk._types import SamplingContext, ProfilerMode
 
     ThreadId = str
 
@@ -148,6 +149,23 @@ def is_gevent():
 PROFILE_MINIMUM_SAMPLES = 2
 
 
+def has_profiling_enabled(options):
+    # type: (Dict[str, Any]) -> bool
+    profiles_sampler = options["profiles_sampler"]
+    if profiles_sampler is not None:
+        return True
+
+    profiles_sample_rate = options["profiles_sample_rate"]
+    if profiles_sample_rate is not None and profiles_sample_rate > 0:
+        return True
+
+    profiles_sample_rate = options["_experiments"].get("profiles_sample_rate")
+    if profiles_sample_rate is not None and profiles_sample_rate > 0:
+        return True
+
+    return False
+
+
 def setup_profiler(options):
     # type: (Dict[str, Any]) -> bool
     global _scheduler
@@ -171,7 +189,13 @@ def setup_profiler(options):
     else:
         default_profiler_mode = ThreadScheduler.mode
 
-    profiler_mode = options["_experiments"].get("profiler_mode", default_profiler_mode)
+    if options.get("profiler_mode") is not None:
+        profiler_mode = options["profiler_mode"]
+    else:
+        profiler_mode = (
+            options.get("_experiments", {}).get("profiler_mode")
+            or default_profiler_mode
+        )
 
     if (
         profiler_mode == ThreadScheduler.mode
@@ -491,7 +515,13 @@ def _set_initial_sampling_decision(self, sampling_context):
             return
 
         options = client.options
-        sample_rate = options["_experiments"].get("profiles_sample_rate")
+
+        if callable(options.get("profiles_sampler")):
+            sample_rate = options["profiles_sampler"](sampling_context)
+        elif options["profiles_sample_rate"] is not None:
+            sample_rate = options["profiles_sample_rate"]
+        else:
+            sample_rate = options["_experiments"].get("profiles_sample_rate")
 
         # The profiles_sample_rate option was not set, so profiling
         # was never enabled.
@@ -502,6 +532,13 @@ def _set_initial_sampling_decision(self, sampling_context):
             self.sampled = False
             return
 
+        if not is_valid_sample_rate(sample_rate, source="Profiling"):
+            logger.warning(
+                "[Profiling] Discarding profile because of invalid sample rate."
+            )
+            self.sampled = False
+            return
+
         # Now we roll the dice. random.random is inclusive of 0, but not of 1,
         # so strict < is safe here. In case sample_rate is a boolean, cast it
         # to a float (True becomes 1.0 and False becomes 0.0)
@@ -695,7 +732,7 @@ def valid(self):
 
 
 class Scheduler(object):
-    mode = "unknown"
+    mode = "unknown"  # type: ProfilerMode
 
     def __init__(self, frequency):
         # type: (int) -> None
@@ -824,7 +861,7 @@ class ThreadScheduler(Scheduler):
     the sampler at a regular interval.
     """
 
-    mode = "thread"
+    mode = "thread"  # type: ProfilerMode
     name = "sentry.profiler.ThreadScheduler"
 
     def __init__(self, frequency):
@@ -905,7 +942,7 @@ class GeventScheduler(Scheduler):
        results in a sample containing only the sampler's code.
     """
 
-    mode = "gevent"
+    mode = "gevent"  # type: ProfilerMode
     name = "sentry.profiler.GeventScheduler"
 
     def __init__(self, frequency):

sentry_sdk/tracing.py

@@ -5,7 +5,7 @@
 
 import sentry_sdk
 from sentry_sdk.consts import INSTRUMENTER
-from sentry_sdk.utils import logger, nanosecond_time
+from sentry_sdk.utils import is_valid_sample_rate, logger, nanosecond_time
 from sentry_sdk._types import TYPE_CHECKING
 
 
@@ -722,7 +722,7 @@ def _set_initial_sampling_decision(self, sampling_context):
         # Since this is coming from the user (or from a function provided by the
         # user), who knows what we might get. (The only valid values are
         # booleans or numbers between 0 and 1.)
-        if not is_valid_sample_rate(sample_rate):
+        if not is_valid_sample_rate(sample_rate, source="Tracing"):
             logger.warning(
                 "[Tracing] Discarding {transaction_description} because of invalid sample rate.".format(
                     transaction_description=transaction_description,
@@ -810,6 +810,5 @@ def finish(self, hub=None, end_timestamp=None):
     EnvironHeaders,
     extract_sentrytrace_data,
     has_tracing_enabled,
-    is_valid_sample_rate,
     maybe_create_breadcrumbs_from_span,
 )

sentry_sdk/tracing_utils.py

@@ -1,17 +1,12 @@
 import re
 import contextlib
-import math
-
-from numbers import Real
-from decimal import Decimal
 
 import sentry_sdk
 from sentry_sdk.consts import OP
 
 from sentry_sdk.utils import (
     capture_internal_exceptions,
     Dsn,
-    logger,
     to_string,
 )
 from sentry_sdk._compat import PY2, iteritems
@@ -100,37 +95,6 @@ def has_tracing_enabled(options):
     )
 
 
-def is_valid_sample_rate(rate):
-    # type: (Any) -> bool
-    """
-    Checks the given sample rate to make sure it is valid type and value (a
-    boolean or a number between 0 and 1, inclusive).
-    """
-
-    # both booleans and NaN are instances of Real, so a) checking for Real
-    # checks for the possibility of a boolean also, and b) we have to check
-    # separately for NaN and Decimal does not derive from Real so need to check that too
-    if not isinstance(rate, (Real, Decimal)) or math.isnan(rate):
-        logger.warning(
-            "[Tracing] Given sample rate is invalid. Sample rate must be a boolean or a number between 0 and 1. Got {rate} of type {type}.".format(
-                rate=rate, type=type(rate)
-            )
-        )
-        return False
-
-    # in case rate is a boolean, it will get cast to 1 if it's True and 0 if it's False
-    rate = float(rate)
-    if rate < 0 or rate > 1:
-        logger.warning(
-            "[Tracing] Given sample rate is invalid. Sample rate must be between 0 and 1. Got {rate}.".format(
-                rate=rate
-            )
-        )
-        return False
-
-    return True
-
-
 @contextlib.contextmanager
 def record_sql_queries(
     hub,  # type: sentry_sdk.Hub

sentry_sdk/utils.py

@@ -2,13 +2,16 @@
 import json
 import linecache
 import logging
+import math
 import os
 import re
 import subprocess
 import sys
 import threading
 import time
 from collections import namedtuple
+from decimal import Decimal
+from numbers import Real
 
 try:
     # Python 3
@@ -1260,6 +1263,37 @@ def parse_url(url, sanitize=True):
     return ParsedUrl(url=base_url, query=parsed_url.query, fragment=parsed_url.fragment)
 
 
+def is_valid_sample_rate(rate, source):
+    # type: (Any, str) -> bool
+    """
+    Checks the given sample rate to make sure it is valid type and value (a
+    boolean or a number between 0 and 1, inclusive).
+    """
+
+    # both booleans and NaN are instances of Real, so a) checking for Real
+    # checks for the possibility of a boolean also, and b) we have to check
+    # separately for NaN and Decimal does not derive from Real so need to check that too
+    if not isinstance(rate, (Real, Decimal)) or math.isnan(rate):
+        logger.warning(
+            "{source} Given sample rate is invalid. Sample rate must be a boolean or a number between 0 and 1. Got {rate} of type {type}.".format(
+                source=source, rate=rate, type=type(rate)
+            )
+        )
+        return False
+
+    # in case rate is a boolean, it will get cast to 1 if it's True and 0 if it's False
+    rate = float(rate)
+    if rate < 0 or rate > 1:
+        logger.warning(
+            "{source} Given sample rate is invalid. Sample rate must be between 0 and 1. Got {rate}.".format(
+                source=source, rate=rate
+            )
+        )
+        return False
+
+    return True
+
+
 if PY37:
 
     def nanosecond_time():