ref(sdk): Remove excessive json.loads spans (#64883)

### Summary
Many of our transactions use up their span limit immediately because of
json.loads creating spans for every field call etc.

![Screenshot 2024-02-08 at 11 25 04
AM](https://github.com/getsentry/sentry/assets/6111995/cc5181cf-845a-4546-9703-8989c68d0a52)

I ran the devserver and tried a couple endpoints and removed spans for
anything I saw spamming. Feel free to re-add it if you find a python
profile for your transaction that shows json.loads is taking up a
significant amount of time, I'm just trying to get most of them that
concern me at the moment without opening a bunch of PRs one by one.

In the future we should probably consider moving these over to a metric
as the span offers no additional information, so we can avoid this
altogether. At the moment, since json can be called at startup there is
currently an issue with config not being loaded so metrics can be used
here so that would need to be resolved first.

src/sentry/db/models/fields/jsonfield.py

@@ -83,8 +83,8 @@ def get_default(self):
             if callable(default):
                 default = default()
             if isinstance(default, str):
-                return json.loads(default)
-            return json.loads(json.dumps(default))
+                return json.loads(default, skip_trace=True)
+            return json.loads(json.dumps(default), skip_trace=True)
         return super().get_default()
 
     def get_internal_type(self):
@@ -101,7 +101,7 @@ def to_python(self, value):
                 if self.blank:
                     return ""
             try:
-                value = json.loads(value)
+                value = json.loads(value, skip_trace=True)
             except ValueError:
                 msg = self.error_messages["invalid"] % value
                 raise ValidationError(msg)

src/sentry/db/models/fields/picklefield.py

@@ -26,6 +26,6 @@ def to_python(self, value):
         if value is None:
             return None
         try:
-            return json.loads(value)
+            return json.loads(value, skip_trace=True)
         except (ValueError, TypeError):
             return super().to_python(value)

src/sentry/ingest/billing_metrics_consumer.py

@@ -85,7 +85,9 @@ def submit(self, message: Message[KafkaPayload]) -> None:
         self.__next_step.submit(message)
 
     def _get_payload(self, message: Message[KafkaPayload]) -> GenericMetric:
-        payload = json.loads(message.payload.value.decode("utf-8"), use_rapid_json=True)
+        payload = json.loads(
+            message.payload.value.decode("utf-8"), use_rapid_json=True, skip_trace=True
+        )
         return cast(GenericMetric, payload)
 
     def _count_processed_items(self, generic_metric: GenericMetric) -> Mapping[DataCategory, int]:

src/sentry/ingest/consumer/processors.py

@@ -106,7 +106,7 @@ def process_event(message: IngestMessage, project: Project) -> None:
     # serializing it again.
     # XXX: Do not use CanonicalKeyDict here. This may break preprocess_event
     # which assumes that data passed in is a raw dictionary.
-    data = json.loads(payload, use_rapid_json=True)
+    data = json.loads(payload, use_rapid_json=True, skip_trace=True)
     if project_id == settings.SENTRY_PROJECT:
         metrics.incr(
             "internal.captured.ingest_consumer.parsed",

src/sentry/utils/codecs.py

@@ -75,7 +75,7 @@ def encode(self, value: JSONData) -> str:
         return str(json.dumps(value))
 
     def decode(self, value: str) -> JSONData:
-        return json.loads(value)
+        return json.loads(value, skip_trace=True)
 
 
 class ZlibCodec(Codec[bytes, bytes]):

src/sentry/utils/snuba.py

@@ -958,7 +958,7 @@ def _bulk_snuba_query(
     for index, item in enumerate(query_results):
         response, _, reverse = item
         try:
-            body = json.loads(response.data)
+            body = json.loads(response.data, skip_trace=True)
             if SNUBA_INFO:
                 if "sql" in body:
                     print(  # NOQA: only prints when an env variable is set