ref: use LegacyTextJSONField for large tables (#97571)

these tables are too large to convert to models.JSONField :(

<!-- Describe your PR here. -->

Author: asottile-sentry

migrations_lockfile.txt

@@ -27,7 +27,7 @@ preprod: 0014_commitcomparisons_fk
 
 replays: 0006_add_bulk_delete_job
 
-sentry: 0966_groupopenperiod_data_pending_inc_detector_id_index
+sentry: 0967_large_tables_legacy_json_field
 
 social_auth: 0003_social_auth_json_field
 

src/sentry/integrations/models/external_issue.py

@@ -8,8 +8,9 @@
 
 from sentry.backup.scopes import RelocationScope
 from sentry.constants import ObjectStatus
-from sentry.db.models import FlexibleForeignKey, JSONField, Model, region_silo_model, sane_repr
+from sentry.db.models import FlexibleForeignKey, Model, region_silo_model, sane_repr
 from sentry.db.models.fields.hybrid_cloud_foreign_key import HybridCloudForeignKey
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.db.models.manager.base import BaseManager
 from sentry.services.eventstore.models import GroupEvent
 
@@ -70,7 +71,7 @@ class ExternalIssue(Model):
     date_added = models.DateTimeField(default=timezone.now)
     title = models.TextField(null=True)
     description = models.TextField(null=True)
-    metadata = JSONField(null=True)
+    metadata = LegacyTextJSONField(default=dict, null=True)
 
     objects: ClassVar[ExternalIssueManager] = ExternalIssueManager()
 

src/sentry/migrations/0967_large_tables_legacy_json_field.py

@@ -0,0 +1,94 @@
+# Generated by Django 5.2.1 on 2025-08-11 14:46
+
+from django.db import migrations
+
+import sentry.db.models.fields.jsonfield
+from sentry.new_migrations.migrations import CheckedMigration
+
+
+class Migration(CheckedMigration):
+    # This flag is used to mark that a migration shouldn't be automatically run in production.
+    # This should only be used for operations where it's safe to run the migration after your
+    # code has deployed. So this should not be used for most operations that alter the schema
+    # of a table.
+    # Here are some things that make sense to mark as post deployment:
+    # - Large data migrations. Typically we want these to be run manually so that they can be
+    #   monitored and not block the deploy for a long period of time while they run.
+    # - Adding indexes to large tables. Since this can take a long time, we'd generally prefer to
+    #   run this outside deployments so that we don't block them. Note that while adding an index
+    #   is a schema change, it's completely safe to run the operation after the code has deployed.
+    # Once deployed, run these manually via: https://develop.sentry.dev/database-migrations/#migration-deployment
+
+    is_post_deployment = False
+
+    dependencies = [
+        ("sentry", "0966_groupopenperiod_data_pending_inc_detector_id_index"),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name="controlfile",
+            name="headers",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="externalissue",
+            name="metadata",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict, null=True),
+        ),
+        migrations.AlterField(
+            model_name="featureadoption",
+            name="data",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="file",
+            name="headers",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="grouphashmetadata",
+            name="hashing_metadata",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(null=True),
+        ),
+        migrations.AlterField(
+            model_name="groupinbox",
+            name="reason_details",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict, null=True),
+        ),
+        migrations.AlterField(
+            model_name="organizationonboardingtask",
+            name="data",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="projectdebugfile",
+            name="data",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict, null=True),
+        ),
+        migrations.AlterField(
+            model_name="projectkey",
+            name="data",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="promptsactivity",
+            name="data",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="pullrequestcomment",
+            name="reactions",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(null=True),
+        ),
+        migrations.AlterField(
+            model_name="release",
+            name="data",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name="repository",
+            name="config",
+            field=sentry.db.models.fields.jsonfield.LegacyTextJSONField(default=dict),
+        ),
+    ]

src/sentry/models/debugfile.py

@@ -27,11 +27,11 @@
 from sentry.db.models import (
     BoundedBigIntegerField,
     FlexibleForeignKey,
-    JSONField,
     Model,
     region_silo_model,
     sane_repr,
 )
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.db.models.manager.base import BaseManager
 from sentry.models.files.file import File
 from sentry.models.files.utils import clear_cached_files
@@ -126,7 +126,7 @@ class ProjectDebugFile(Model):
     project_id = BoundedBigIntegerField(null=True, db_index=True)
     debug_id = models.CharField(max_length=64, db_column="uuid")
     code_id = models.CharField(max_length=64, null=True)
-    data: models.Field[dict[str, Any] | None, dict[str, Any] | None] = JSONField(null=True)
+    data = LegacyTextJSONField(default=dict, null=True)
     date_accessed = models.DateTimeField(default=timezone.now, db_default=Now())
 
     objects: ClassVar[ProjectDebugFileManager] = ProjectDebugFileManager()

src/sentry/models/featureadoption.py

@@ -10,7 +10,8 @@
 from sentry.adoption import manager
 from sentry.adoption.manager import UnknownFeature
 from sentry.backup.scopes import RelocationScope
-from sentry.db.models import FlexibleForeignKey, JSONField, Model, region_silo_model, sane_repr
+from sentry.db.models import FlexibleForeignKey, Model, region_silo_model, sane_repr
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.db.models.manager.base import BaseManager
 from sentry.utils.redis import (
     get_dynamic_cluster_from_options,
@@ -235,7 +236,7 @@ class FeatureAdoption(Model):
     date_completed = models.DateTimeField(default=timezone.now)
     complete = models.BooleanField(default=False)
     applicable = models.BooleanField(default=True)  # Is this feature applicable to this team?
-    data = JSONField()
+    data = LegacyTextJSONField(default=dict)
 
     objects: ClassVar[FeatureAdoptionManager] = FeatureAdoptionManager()
 

src/sentry/models/files/abstractfile.py

@@ -19,7 +19,8 @@
 
 from sentry.backup.scopes import RelocationScope
 from sentry.celery import SentryTask
-from sentry.db.models import JSONField, Model, WrappingU32IntegerField
+from sentry.db.models import Model, WrappingU32IntegerField
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.models.files.abstractfileblob import AbstractFileBlob
 from sentry.models.files.abstractfileblobindex import AbstractFileBlobIndex
 from sentry.models.files.utils import DEFAULT_BLOB_SIZE, AssembleChecksumMismatch, nooplogger
@@ -219,7 +220,7 @@ class AbstractFile(Model, _Parent[BlobIndexType, BlobType]):
     name = models.TextField()
     type = models.CharField(max_length=64)
     timestamp = models.DateTimeField(default=timezone.now, db_index=True)
-    headers: models.Field[dict[str, Any], dict[str, Any]] = JSONField()
+    headers = LegacyTextJSONField(default=dict)
     size = WrappingU32IntegerField(null=True)
     checksum = models.CharField(max_length=40, null=True, db_index=True)
 

src/sentry/models/grouphashmetadata.py

@@ -8,8 +8,7 @@
 from sentry.db.models import Model, region_silo_model
 from sentry.db.models.base import sane_repr
 from sentry.db.models.fields.foreignkey import FlexibleForeignKey
-from sentry.db.models.fields.jsonfield import JSONField
-from sentry.types.grouphash_metadata import HashingMetadata
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 
 # The current version of the metadata schema. Any record encountered whose schema version is earlier
 # than this will have its data updated and its version set to this value. Stored as a string for
@@ -96,9 +95,7 @@ class GroupHashMetadata(Model):
     # Metadata about the inputs to the hashing process and the hashing process itself (what
     # fingerprinting rules were matched? did we parameterize the message? etc.). For the specific
     # data stored, see the class definitions of the `HashingMetadata` subtypes.
-    hashing_metadata: models.Field[HashingMetadata | None, HashingMetadata | None] = JSONField(
-        null=True
-    )
+    hashing_metadata = LegacyTextJSONField(null=True)
 
     # SEER
 

src/sentry/models/groupinbox.py

@@ -10,7 +10,8 @@
 from django.utils import timezone
 
 from sentry.backup.scopes import RelocationScope
-from sentry.db.models import FlexibleForeignKey, JSONField, Model, region_silo_model
+from sentry.db.models import FlexibleForeignKey, Model, region_silo_model
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.db.models.manager.base_query_set import BaseQuerySet
 from sentry.models.activity import Activity
 from sentry.models.group import Group
@@ -57,7 +58,7 @@ class GroupInbox(Model):
     project = FlexibleForeignKey("sentry.Project", null=True, db_constraint=False)
     organization = FlexibleForeignKey("sentry.Organization", null=True, db_constraint=False)
     reason = models.PositiveSmallIntegerField(null=False, default=GroupInboxReason.NEW.value)
-    reason_details = JSONField(null=True)
+    reason_details = LegacyTextJSONField(default=dict, null=True)
     date_added = models.DateTimeField(default=timezone.now, db_index=True)
 
     class Meta:
@@ -71,9 +72,6 @@ def add_group_to_inbox(
     reason: GroupInboxReason,
     reason_details: InboxReasonDetails | None = None,
 ) -> GroupInbox:
-    if reason_details is not None and reason_details["until"] is not None:
-        reason_details["until"] = reason_details["until"].replace(microsecond=0)
-
     group_inbox, _ = GroupInbox.objects.get_or_create(
         group=group,
         defaults={
@@ -137,7 +135,7 @@ def bulk_remove_groups_from_inbox(
 
 
 class InboxReasonDetails(TypedDict):
-    until: datetime | None
+    until: str | None  # datetime str
     count: int | None
     window: int | None
     user_count: int | None

src/sentry/models/organizationonboardingtask.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import enum
-from typing import Any, ClassVar
+from typing import ClassVar
 
 from django.conf import settings
 from django.core.cache import cache
@@ -13,12 +13,12 @@
 from sentry.db.models import (
     BoundedPositiveIntegerField,
     FlexibleForeignKey,
-    JSONField,
     Model,
     region_silo_model,
     sane_repr,
 )
 from sentry.db.models.fields.hybrid_cloud_foreign_key import HybridCloudForeignKey
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.db.models.manager.base import BaseManager
 
 
@@ -111,7 +111,7 @@ class AbstractOnboardingTask(Model):
         "sentry.Project", db_constraint=False, null=True, on_delete=SET_NULL
     )
     # INVITE_MEMBER { invited_member: user.id }
-    data: models.Field[dict[str, Any], dict[str, Any]] = JSONField()
+    data = LegacyTextJSONField(default=dict)
 
     # abstract
     TASK_LOOKUP_BY_KEY: dict[str, int]

src/sentry/models/projectkey.py

@@ -3,7 +3,7 @@
 import enum
 import re
 import secrets
-from typing import Any, ClassVar
+from typing import ClassVar
 from urllib.parse import urlparse
 
 import petname
@@ -22,11 +22,11 @@
 from sentry.db.models import (
     BoundedPositiveIntegerField,
     FlexibleForeignKey,
-    JSONField,
     Model,
     region_silo_model,
     sane_repr,
 )
+from sentry.db.models.fields.jsonfield import LegacyTextJSONField
 from sentry.db.models.manager.base import BaseManager
 from sentry.silo.base import SiloMode
 from sentry.tasks.relay import schedule_invalidate_project_config
@@ -120,7 +120,7 @@ class roles(TypedClassBitField):
         cache_ttl=60 * 30,
     )
 
-    data: models.Field[dict[str, Any], dict[str, Any]] = JSONField()
+    data = LegacyTextJSONField(default=dict)
 
     use_case = models.CharField(
         max_length=32,