Merge branch 'master' into hogql-calculation-bis

.github/workflows/ci-backend.yml

@@ -39,7 +39,7 @@ jobs:
     # Job to decide if we should run backend ci
     # See https://github.com/dorny/paths-filter#conditional-execution for more details
     changes:
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         timeout-minutes: 5
         if: github.repository == 'PostHog/posthog'
         name: Determine need to run backend checks
@@ -90,7 +90,7 @@ jobs:
         timeout-minutes: 30
 
         name: Python code quality checks
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
 
         steps:
             # If this run wasn't initiated by the bot (meaning: snapshot update) and we've determined
@@ -171,7 +171,7 @@ jobs:
         timeout-minutes: 10
 
         name: Validate Django and CH migrations
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
 
         steps:
             - uses: actions/checkout@v3
@@ -236,7 +236,7 @@ jobs:
         timeout-minutes: 30
 
         name: Django tests – ${{ matrix.segment }} (persons-on-events ${{ matrix.person-on-events && 'on' || 'off' }}), Py ${{ matrix.python-version }}, ${{ matrix.clickhouse-server-image }} (${{matrix.group}}/${{ matrix.concurrency }})
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
 
         strategy:
             fail-fast: false
@@ -319,7 +319,7 @@ jobs:
                 clickhouse-server-image:
                     ['clickhouse/clickhouse-server:23.11.2.11-alpine', 'clickhouse/clickhouse-server:23.12.5.81-alpine']
         if: needs.changes.outputs.backend == 'true'
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         steps:
             - name: 'Checkout repo'
               uses: actions/checkout@v3
@@ -373,7 +373,7 @@ jobs:
     calculate-running-time:
         name: Calculate running time
         needs: [django, async-migrations]
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         if: needs.changes.outputs.backend == 'true'
         steps:
             - name: Calculate running time

.github/workflows/ci-e2e.yml

@@ -14,7 +14,7 @@ concurrency:
 
 jobs:
     changes:
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         timeout-minutes: 5
         if: github.repository == 'PostHog/posthog'
         name: Determine need to run E2E checks
@@ -55,11 +55,10 @@ jobs:
     chunks:
         needs: changes
         name: Cypress preparation
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         timeout-minutes: 5
         outputs:
             chunks: ${{ steps.chunk.outputs.chunks }}
-
         steps:
             - name: Check out
               uses: actions/checkout@v3
@@ -70,7 +69,7 @@ jobs:
 
     container:
         name: Build and cache container image
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         timeout-minutes: 60
         needs: [changes]
         permissions:
@@ -94,7 +93,7 @@ jobs:
 
     cypress:
         name: Cypress E2E tests (${{ strategy.job-index }})
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         timeout-minutes: 60
         needs: [chunks, changes, container]
         permissions:
@@ -279,7 +278,7 @@ jobs:
 
     calculate-running-time:
         name: Calculate running time
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         needs: [cypress]
         if: needs.changes.outputs.shouldTriggerCypress == 'true'
         steps:

.github/workflows/ci-plugin-server.yml

@@ -24,7 +24,7 @@ jobs:
     # Job to decide if we should run plugin server ci
     # See https://github.com/dorny/paths-filter#conditional-execution for more details
     changes:
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         timeout-minutes: 5
         if: github.repository == 'PostHog/posthog'
         name: Determine need to run plugin server checks
@@ -54,7 +54,7 @@ jobs:
         name: Code quality
         needs: changes
         if: needs.changes.outputs.plugin-server == 'true'
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
         defaults:
             run:
                 working-directory: 'plugin-server'
@@ -85,7 +85,7 @@ jobs:
     tests:
         name: Plugin Server Tests (${{matrix.shard}})
         needs: changes
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
 
         strategy:
             fail-fast: false
@@ -180,7 +180,7 @@ jobs:
         name: Functional tests (POE=${{matrix.POE_EMBRACE_JOIN_FOR_TEAMS}},RDK=${{matrix.KAFKA_CONSUMPTION_USE_RDKAFKA}})
         needs: changes
         if: needs.changes.outputs.plugin-server == 'true'
-        runs-on: depot-ubuntu-latest-4
+        runs-on: ubuntu-latest
 
         strategy:
             fail-fast: false

.vscode/launch.json

@@ -72,8 +72,7 @@
                 "DATABASE_URL": "postgres://posthog:posthog@localhost:5432/posthog",
                 "SKIP_SERVICE_VERSION_REQUIREMENTS": "1",
                 "PRINT_SQL": "1",
-                "BILLING_SERVICE_URL": "https://billing.dev.posthog.dev",
-                "RECORDINGS_INGESTER_URL": "http://localhost:6738"
+                "BILLING_SERVICE_URL": "https://billing.dev.posthog.dev"
             },
             "console": "integratedTerminal",
             "python": "${workspaceFolder}/env/bin/python",

bin/start

@@ -7,7 +7,6 @@ trap "trap - SIGTERM && kill -- -$$" SIGINT SIGTERM EXIT
 export DEBUG=${DEBUG:-1}
 export SKIP_SERVICE_VERSION_REQUIREMENTS=1
 export BILLING_SERVICE_URL=${BILLING_SERVICE_URL:-https://billing.dev.posthog.dev}
-export RECORDINGS_INGESTER_URL=${RECORDINGS_INGESTER_URL:-http://localhost:6738}
 
 service_warning() {
   echo -e "\033[0;31m$1 isn't ready. You can run the stack with:\ndocker compose -f docker-compose.dev.yml up\nIf you have already ran that, just make sure that services are starting properly, and sit back.\nWaiting for $1 to start...\033[0m"

docker-compose.base.yml

@@ -151,6 +151,7 @@ services:
         volumes:
             - /var/lib/elasticsearch/data
     temporal:
+        restart: on-failure
         environment:
             - DB=postgresql
             - DB_PORT=5432
@@ -169,8 +170,11 @@ services:
         volumes:
             - ./docker/temporal/dynamicconfig:/etc/temporal/config/dynamicconfig
         depends_on:
-            - db
-            - elasticsearch
+            db:
+                condition: service_healthy
+            elasticsearch:
+                condition: service_started
+
     temporal-admin-tools:
         environment:
             - TEMPORAL_CLI_ADDRESS=temporal:7233

docker-compose.dev.yml

@@ -121,6 +121,7 @@ services:
         extends:
             file: docker-compose.base.yml
             service: temporal
+
     temporal-admin-tools:
         extends:
             file: docker-compose.base.yml

docker-compose.hobby.yml

@@ -76,7 +76,6 @@ services:
             SENTRY_DSN: $SENTRY_DSN
             SITE_URL: https://$DOMAIN
             SECRET_KEY: $POSTHOG_SECRET
-            RECORDINGS_INGESTER_URL: http://plugins:6738
         depends_on:
             - db
             - redis

ee/clickhouse/views/experiments.py

@@ -18,6 +18,7 @@
     ClickhouseTrendExperimentResult,
 )
 from ee.clickhouse.queries.experiments.utils import requires_flag_warning
+from posthog.api.cohort import CohortSerializer
 from posthog.api.feature_flag import FeatureFlagSerializer, MinimalFeatureFlagSerializer
 from posthog.api.routing import TeamAndOrgViewSetMixin
 from posthog.api.shared import UserBasicSerializer
@@ -150,6 +151,7 @@ class Meta:
             "end_date",
             "feature_flag_key",
             "feature_flag",
+            "exposure_cohort",
             "parameters",
             "secondary_metrics",
             "filters",
@@ -164,6 +166,7 @@ class Meta:
             "created_at",
             "updated_at",
             "feature_flag",
+            "exposure_cohort",
         ]
 
     def validate_parameters(self, value):
@@ -353,3 +356,89 @@ def requires_flag_implementation(self, request: Request, *args: Any, **kwargs: A
         warning = requires_flag_warning(filter, self.team)
 
         return Response({"result": warning})
+
+    @action(methods=["POST"], detail=True)
+    def create_exposure_cohort_for_experiment(self, request: Request, *args: Any, **kwargs: Any) -> Response:
+        experiment = self.get_object()
+        flag = getattr(experiment, "feature_flag", None)
+        if not flag:
+            raise ValidationError("Experiment does not have a feature flag")
+
+        if not experiment.start_date:
+            raise ValidationError("Experiment does not have a start date")
+
+        if experiment.exposure_cohort:
+            raise ValidationError("Experiment already has an exposure cohort")
+
+        exposure_filter_data = (experiment.parameters or {}).get("custom_exposure_filter")
+        exposure_filter = None
+        if exposure_filter_data:
+            exposure_filter = Filter(data={**exposure_filter_data, "is_simplified": True}, team=experiment.team)
+
+        target_entity: int | str = "$feature_flag_called"
+        target_entity_type = "events"
+        target_filters = [
+            {
+                "key": "$feature_flag",
+                "value": [flag.key],
+                "operator": "exact",
+                "type": "event",
+            }
+        ]
+
+        if exposure_filter:
+            entity = exposure_filter.entities[0]
+            if entity.id:
+                target_entity_type = entity.type if entity.type in ["events", "actions"] else "events"
+                target_entity = entity.id
+                if entity.type == "actions":
+                    try:
+                        target_entity = int(target_entity)
+                    except ValueError:
+                        raise ValidationError("Invalid action ID")
+
+                target_filters = [
+                    prop.to_dict()
+                    for prop in entity.property_groups.flat
+                    if prop.type in ("event", "feature", "element", "hogql")
+                ]
+
+        cohort_serializer = CohortSerializer(
+            data={
+                "is_static": False,
+                "name": f'Users exposed to experiment "{experiment.name}"',
+                "is_calculating": True,
+                "filters": {
+                    "properties": {
+                        "type": "OR",
+                        "values": [
+                            {
+                                "type": "OR",
+                                "values": [
+                                    {
+                                        "type": "behavioral",
+                                        "value": "performed_event",
+                                        "key": target_entity,
+                                        "negation": False,
+                                        "event_type": target_entity_type,
+                                        "event_filters": target_filters,
+                                        "explicit_datetime": experiment.start_date.isoformat(),
+                                    }
+                                ],
+                            }
+                        ],
+                    }
+                },
+            },
+            context={
+                "request": request,
+                "team": self.team,
+                "team_id": self.team_id,
+            },
+        )
+
+        cohort_serializer.is_valid(raise_exception=True)
+        cohort = cohort_serializer.save()
+        experiment.exposure_cohort = cohort
+        experiment.save(update_fields=["exposure_cohort"])
+        return Response({"cohort": cohort_serializer.data}, status=201)

ee/clickhouse/views/test/__snapshots__/test_clickhouse_experiment_secondary_results.ambr

@@ -1,7 +1,7 @@
 # serializer version: 1
 # name: ClickhouseTestExperimentSecondaryResults.test_basic_secondary_metric_results
   '''
-  /* user_id:109 celery:posthog.tasks.tasks.sync_insight_caching_state */
+  /* user_id:106 celery:posthog.tasks.tasks.sync_insight_caching_state */
   SELECT team_id,
          date_diff('second', max(timestamp), now()) AS age
   FROM events