BerriAI/litellm · error · Exception
Error updating policy in DB: {e}
Error message
Error updating policy in DB: {e} What it means
Catch-all from the policy update routine: building the update payload (including GuardrailPipeline validation and JSON serialization of condition/pipeline) or executing the Prisma update failed; the original exception is embedded in the message.
Source
Thrown at litellm/proxy/policy_engine/policy_registry.py:492
if policy_request.guardrails_remove is not None:
update_data["guardrails_remove"] = policy_request.guardrails_remove
if policy_request.condition is not None:
update_data["condition"] = json.dumps(policy_request.condition.model_dump())
if policy_request.pipeline is not None:
validated_pipeline: Final = GuardrailPipeline(**policy_request.pipeline)
update_data["pipeline"] = json.dumps(validated_pipeline.model_dump())
updated_policy: Final = await _policy_table(prisma_client).update(
where={"policy_id": policy_id},
data=update_data,
)
# Do NOT update in-memory registry: drafts are not loaded into memory.
return _row_to_policy_db_response(updated_policy)
except Exception as e:
verbose_proxy_logger.exception("Error updating policy in DB: %s", e)
raise Exception(f"Error updating policy in DB: {e}")
async def delete_policy_from_db(
self,
policy_id: str,
prisma_client: "PrismaClient",
) -> Mapping[str, str]:
"""
Delete a policy version from the database.
If the deleted version was production, it is removed from the in-memory
registry. No other version is auto-promoted; admin must explicitly promote.
Args:
policy_id: The ID of the policy version to delete
prisma_client: The Prisma client instance
Returns:
Dict with "message" and optional "warning" if production was deleted.View on GitHub (pinned to 77b7c6c40c)
Solutions
- Check DB connectivity and logs for the underlying error, then retry the update.
Defensive patterns
Strategy: try-catch
When it happens
Trigger: Thrown at litellm/proxy/policy_engine/policy_registry.py:492 when the library encounters an invalid state.
Common situations: See trigger scenarios.
AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18).
Data as JSON: /api/errors/bbf6fbcb73066a77.
Report an issue: GitHub.