@@ -35,7 +35,7 @@ replays: 0001_squashed_0007_organizationmember_replay_access
 
 seer: 0038_add_workflow_config_scope
 
-sentry: 1230_add_commit_new_repository_id_no_check
+sentry: 1231_drop_group_logger_indexes
 
 social_auth: 0001_squashed_0003_social_auth_json_field
 
 
@@ -0,0 +1,76 @@
+# Generated by Django 5.2.17 on 2026-10-21 13:00
+
+from django.db import migrations, models
+
+from sentry.new_migrations.migrations import CheckedMigration
+from sentry.new_migrations.monkey.special import SafeRunSQL
+
+# Created by `db_index=True` on `Group.logger`: a btree index plus a `varchar_pattern_ops` index
+# for `LIKE` lookups. Nothing queries `Group` by `logger` in Postgres.
+# `IF EXISTS` keeps these drops a no-op on any database where an index is absent.
+LOGGER_INDEX = "sentry_groupedmessage_logger_993cb6d5"
+LOGGER_LIKE_INDEX = "sentry_groupedmessage_logger_993cb6d5_like"
+# South-era names for the same indexes, which may still exist on databases created before Django
+# migrations.
+LEGACY_LOGGER_INDEX = "sentry_groupedmessage_logger"
+LEGACY_LOGGER_LIKE_INDEX = "sentry_groupedmessage_logger_like"
+
+
+def drop_index(name: str, reverse_sql: str) -> SafeRunSQL:
+    return SafeRunSQL(
+        f'DROP INDEX CONCURRENTLY IF EXISTS "{name}"',
+        reverse_sql=reverse_sql,
+        hints={"tables": ["sentry_groupedmessage"]},
+        # The drop waits for open transactions on a hot table, which can outlast the statement
+        # timeout.
+        use_statement_timeout=False,
+    )
+
+
+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 = True
+
+    dependencies = [
+        ("sentry", "1230_add_commit_new_repository_id_no_check"),
+    ]
+
+    operations = [
+        migrations.SeparateDatabaseAndState(
+            # Drop the indexes by their exact names rather than letting `AlterField` introspect
+            # and drop every single-column index on `logger`.
+            database_operations=[
+                drop_index(
+                    LOGGER_INDEX,
+                    f'CREATE INDEX CONCURRENTLY IF NOT EXISTS "{LOGGER_INDEX}" '
+                    'ON "sentry_groupedmessage" ("logger")',
+                ),
+                drop_index(
+                    LOGGER_LIKE_INDEX,
+                    f'CREATE INDEX CONCURRENTLY IF NOT EXISTS "{LOGGER_LIKE_INDEX}" '
+                    'ON "sentry_groupedmessage" ("logger" varchar_pattern_ops)',
+                ),
+                # Reversing the drops above recreates the indexes under their current names.
+                drop_index(LEGACY_LOGGER_INDEX, migrations.RunSQL.noop),
+                drop_index(LEGACY_LOGGER_LIKE_INDEX, migrations.RunSQL.noop),
+            ],
+            state_operations=[
+                migrations.AlterField(
+                    model_name="group",
+                    name="logger",
+                    field=models.CharField(blank=True, default="", max_length=64),
+                ),
+            ],
+        ),
+    ]
@@ -864,9 +864,7 @@ class Group(Model):
     __relocation_scope__ = RelocationScope.Excluded
 
     project = FlexibleForeignKey("sentry.Project")
-    logger = models.CharField(
-        max_length=64, blank=True, default=str(DEFAULT_LOGGER_NAME), db_index=True
-    )
+    logger = models.CharField(max_length=64, blank=True, default=str(DEFAULT_LOGGER_NAME))
     level = BoundedPositiveIntegerField(
         choices=[(key, str(val)) for key, val in sorted(LOG_LEVELS.items())],
         default=logging.ERROR,