Telemetry: delete old BuildData models

readthedocs/settings/base.py

@@ -124,6 +124,9 @@ def SESSION_COOKIE_SAMESITE(self):
     RTD_ANALYTICS_DEFAULT_RETENTION_DAYS = 30 * 3
     RTD_AUDITLOGS_DEFAULT_RETENTION_DAYS = 30 * 3
 
+    # Keep BuildData models on database during this time
+    RTD_TELEMETRY_DATA_RETENTION_DAYS = 30 * 6  # 180 days / 6 months
+
     # Database and API hitting settings
     DONT_HIT_API = False
     DONT_HIT_DB = True
@@ -419,6 +422,11 @@ def TEMPLATES(self):
             'schedule': crontab(minute=0, hour=1),
             'options': {'queue': 'web'},
         },
+        'every-day-delete-old-buildata-models': {
+            'task': 'readthedocs.telemetry.tasks.delete_old_build_data',
+            'schedule': crontab(minute=0, hour=2),
+            'options': {'queue': 'web'},
+        },
         'every-day-resync-sso-organization-users': {
             'task': 'readthedocs.oauth.tasks.sync_remote_repositories_organizations',
             'schedule': crontab(minute=0, hour=4),

readthedocs/telemetry/tasks.py

@@ -1,5 +1,8 @@
 """Tasks related to telemetry."""
 
+from django.conf import settings
+from django.utils import timezone
+
 from readthedocs.builds.models import Build
 from readthedocs.telemetry.models import BuildData
 from readthedocs.worker import app
@@ -16,3 +19,18 @@ def save_build_data(build_id, data):
     build = Build.objects.filter(id=build_id).first()
     if build:
         BuildData.objects.collect(build, data)
+
+
+@app.task(queue="web")
+def delete_old_build_data():
+    """
+    Delete BuildData models older than ``RTD_TELEMETRY_DATA_RETENTION_DAYS``.
+
+    This is intended to run from a periodic task daily.
+
+    NOTE: the logic of this task could be improved to keep longer data we care
+          more (eg. active projects )and remove data we don't (eg. builds from spam projects)
+    """
+    retention_days = settings.RTD_TELEMETRY_DATA_RETENTION_DAYS
+    days_ago = timezone.now().date() - timezone.timedelta(days=retention_days)
+    return BuildData.objects.filter(created__lt=days_ago).delete()