feat(tags): Adds sampling to project tag keys endpoint (#87985)

Updates the project `/tags` endpoint to clamp timerange and sample
events when feature flagged with `organizations:tag-key-sample-n`.
These changes were copied over from the organization level `/tags`
endpoint

Author: edwardgou-sentry

src/sentry/api/endpoints/project_tags.py

@@ -1,11 +1,14 @@
+import datetime
+
 from rest_framework.request import Request
 from rest_framework.response import Response
 
-from sentry import tagstore
+from sentry import features, options, tagstore
 from sentry.api.api_owners import ApiOwner
 from sentry.api.api_publish_status import ApiPublishStatus
 from sentry.api.base import EnvironmentMixin, region_silo_endpoint
 from sentry.api.bases.project import ProjectEndpoint
+from sentry.api.utils import clamp_date_range, default_start_end_dates
 from sentry.constants import DS_DENYLIST, PROTECTED_TAG_KEYS
 from sentry.models.environment import Environment
 
@@ -23,7 +26,7 @@ def get(self, request: Request, project) -> Response:
         except Environment.DoesNotExist:
             tag_keys = []
         else:
-            kwargs = {}
+            kwargs: dict = {}
             if request.GET.get("onlySamplingTags") == "1":
                 kwargs["denylist"] = DS_DENYLIST
 
@@ -38,6 +41,15 @@ def get(self, request: Request, project) -> Response:
 
             include_values_seen = request.GET.get("includeValuesSeen") != "0"
 
+            if features.has("organizations:tag-key-sample-n", project.organization):
+                # Tag queries longer than 14 days tend to time out for large customers. For getting a list of tags, clamping to 14 days is a reasonable compromise of speed vs. completeness
+                (start, end) = clamp_date_range(
+                    default_start_end_dates(),
+                    datetime.timedelta(days=options.get("visibility.tag-key-max-date-range.days")),
+                )
+                kwargs["start"] = start
+                kwargs["end"] = end
+
             tag_keys = sorted(
                 backend.get_tag_keys(
                     project.id,

src/sentry/tagstore/snuba/backend.py

@@ -236,6 +236,14 @@ def __get_tag_keys(
         tenant_ids=None,
         **kwargs,
     ):
+        optimize_kwargs: _OptimizeKwargs = {}
+        if turbo := kwargs.get("turbo"):
+            if isinstance(turbo, bool):
+                optimize_kwargs["turbo"] = turbo
+        if sample := kwargs.get("sample"):
+            if isinstance(sample, int):
+                optimize_kwargs["sample"] = sample
+
         return self.__get_tag_keys_for_projects(
             get_project_list(project_id),
             group,
@@ -248,6 +256,7 @@ def __get_tag_keys(
             dataset=dataset,
             denylist=denylist,
             tenant_ids=tenant_ids,
+            **optimize_kwargs,
         )
 
     def __get_tag_keys_for_projects(
@@ -404,15 +413,30 @@ def get_tag_keys(
         include_values_seen=False,
         denylist=None,
         tenant_ids=None,
+        **kwargs,
     ):
         assert status is TagKeyStatus.ACTIVE
+
+        optimize_kwargs: _OptimizeKwargs = {}
+
+        organization_id = get_organization_id_from_project_ids(get_project_list(project_id))
+        organization = Organization.objects.get_from_cache(id=organization_id)
+        if features.has("organizations:tag-key-sample-n", organization):
+            # Add static sample amount to the query. Turbo will sample at 10% by
+            # default, but organizations with many events still get timeouts. A
+            # static sample creates more consistent performance.
+            optimize_kwargs["turbo"] = True
+            optimize_kwargs["sample"] = options.get("visibility.tag-key-sample-size")
+
         return self.__get_tag_keys(
             project_id,
             None,
             environment_id and [environment_id],
             denylist=denylist,
             tenant_ids=tenant_ids,
             include_values_seen=include_values_seen,
+            **optimize_kwargs,
+            **kwargs,
         )
 
     def get_tag_keys_for_projects(