refactor a bit

Author: Andrii Soldatenko

src/sentry/conf/server.py

@@ -576,7 +576,6 @@ def SOCIAL_AUTH_DEFAULT_USERNAME():
     "sentry.tasks.deletion",
     "sentry.tasks.deliver_from_outbox",
     "sentry.tasks.digests",
-    "sentry.tasks.dynamic_sampling",
     "sentry.tasks.email",
     "sentry.tasks.files",
     "sentry.tasks.groupowner",
@@ -602,6 +601,7 @@ def SOCIAL_AUTH_DEFAULT_USERNAME():
     "sentry.tasks.user_report",
     "sentry.profiles.task",
     "sentry.release_health.tasks",
+    "sentry.dynamic_sampling.tasks",
     "sentry.utils.suspect_resolutions.get_suspect_resolutions",
     "sentry.utils.suspect_resolutions_releases.get_suspect_resolutions_releases",
     "sentry.tasks.derive_code_mappings",

src/sentry/dynamic_sampling/prioritise_projects.py

@@ -4,27 +4,11 @@
 from datetime import datetime, timedelta
 from typing import Mapping, Sequence
 
-from snuba_sdk import (
-    Column,
-    Condition,
-    Direction,
-    Entity,
-    Function,
-    Granularity,
-    Op,
-    OrderBy,
-    Query,
-    Request,
-)
+from snuba_sdk import Column, Condition, Direction, Entity, Granularity, Op, OrderBy, Query, Request
 
-from sentry.release_health.release_monitor.base import BaseReleaseMonitorBackend, Totals
-from sentry.sentry_metrics import indexer
-from sentry.sentry_metrics.configuration import UseCaseKey
-from sentry.sentry_metrics.indexer.strings import SESSION_METRIC_NAMES
-from sentry.sentry_metrics.utils import resolve_tag_key
+from sentry.sentry_metrics.indexer.strings import TRANSACTION_METRICS_NAMES
 from sentry.snuba.dataset import Dataset, EntityKey
-from sentry.snuba.metrics.naming_layer.mri import SessionMRI
-from sentry.utils import metrics
+from sentry.snuba.metrics.naming_layer.mri import TransactionMRI
 from sentry.utils.snuba import raw_snql_query
 
 logger = logging.getLogger(__name__)
@@ -33,7 +17,6 @@
 
 
 def fetch_projects_with_total_volumes() -> Mapping[int, Sequence[int]]:
-    # TODO: (andrii) include only "disconnected" projects or independent in tracing context
     aggregated_projects = defaultdict(list)
     start_time = time.time()
     offset = 0
@@ -52,7 +35,7 @@ def fetch_projects_with_total_volumes() -> Mapping[int, Sequence[int]]:
                     Condition(
                         Column("metric_id"),
                         Op.EQ,
-                        SESSION_METRIC_NAMES[SessionMRI.SESSION.value],
+                        TRANSACTION_METRICS_NAMES[TransactionMRI.COUNT_PER_ROOT_PROJECT.value],
                     ),
                 ],
                 granularity=Granularity(3600),
@@ -66,9 +49,8 @@ def fetch_projects_with_total_volumes() -> Mapping[int, Sequence[int]]:
         )
         request = Request(dataset=Dataset.Metrics.value, app_id="dynamic_sampling", query=query)
         data = raw_snql_query(
-            # TODO: replace to new referrer
             request,
-            referrer="dynamic_sampling.fetch_projects_with_recent_sessions",
+            referrer="dynamic_sampling.fetch_projects_with_total_volumes",
         )["data"]
         count = len(data)
         more_results = count > CHUNK_SIZE
@@ -90,3 +72,7 @@ def fetch_projects_with_total_volumes() -> Mapping[int, Sequence[int]]:
         )
 
     return aggregated_projects
+
+
+def process_projects_with_total_volumes(project_ids):
+    ...

src/sentry/dynamic_sampling/tasks.py

@@ -1,19 +1,36 @@
 import logging
 
+from sentry.dynamic_sampling.prioritise_projects import fetch_projects_with_total_volumes
 from sentry.tasks.base import instrumented_task
 
 CHUNK_SIZE = 1000
 MAX_SECONDS = 60
 
-logger = logging.getLogger("sentry.tasks.dynamic_sampling")
+logger = logging.getLogger(__name__)
 
 
 @instrumented_task(
-    name="sentry.dynamic_sampling.tasks.foo",
-    queue="releasemonitor",
+    name="sentry.dynamic_sampling.tasks.prioritise_projects",
+    queue="dynamicsampling",
     default_retry_delay=5,
     max_retries=5,
 )  # type: ignore
-def foo(**kwargs) -> None:
+def prioritise_projects(**kwargs) -> None:
     for org_id, project_ids in fetch_projects_with_total_volumes().items():
-        process_projects_with_sessions.delay(org_id, project_ids)
+        process_projects_sample_rates.delay(org_id, project_ids)
+
+
+@instrumented_task(
+    name="sentry.dynamic_sampling.process_projects_sample_rates",
+    queue="dynamicsampling",
+    default_retry_delay=5,
+    max_retries=5,
+)  # type: ignore
+def process_projects_sample_rates(org_id, project_ids) -> None:
+    """
+    Takes a single org id and a list of project ids
+    """
+    ...
+
+    # Get adjusted sample rate via adjustment model
+    #

src/sentry/relay/config/__init__.py

@@ -493,6 +493,7 @@ def _filter_option_to_config_setting(flt: _FilterSpec, setting: str) -> Mapping[
     [
         "s:transactions/user@none",
         "d:transactions/duration@millisecond",
+        "c:transactions/count_per_root_project@none",
     ]
 )
 

src/sentry/snuba/metrics/naming_layer/mri.py

@@ -81,6 +81,7 @@ class TransactionMRI(Enum):
     # Ingested
     USER = "s:transactions/user@none"
     DURATION = "d:transactions/duration@millisecond"
+    COUNT_PER_ROOT_PROJECT = "c:transactions/count_per_root_project@none"
     MEASUREMENTS_FCP = "d:transactions/measurements.fcp@millisecond"
     MEASUREMENTS_LCP = "d:transactions/measurements.lcp@millisecond"
     MEASUREMENTS_APP_START_COLD = "d:transactions/measurements.app_start_cold@millisecond"

tests/sentry/dynamic_sampling/test_tasks.py

@@ -0,0 +1,16 @@
+import pytest
+
+from sentry.dynamic_sampling.tasks import fetch_projects_with_total_volumes
+
+
+@pytest.mark.django_db
+def test_simple(default_project):
+    test_data = [
+        {
+            "org_id": [default_project.organization.id],
+            "project_id": [default_project.id],
+        },
+    ]
+    assert 1 == 1
+    _ = test_data
+    fetch_projects_with_total_volumes()