From 15dfaed4627fc82fd60ae89e5974e852c2feadd9 Mon Sep 17 00:00:00 2001
From: Will Miao <willmiao.dev@gmail.com>
Date: Fri, 22 May 2026 07:05:06 +0800
Subject: [PATCH] fix(api): treat transient server errors (524/5xx) as
 non-fatal in model updates (#935)

Teach CivitaiClient.get_model_versions() to recognise Cloudflare 524, generic
5xx, and connection-level errors as transient failures and return None
instead of raising RuntimeError, so a single upstream glitch does not
block the entire batch update or produce a scary traceback.

Also downgrade the generic except Exception log level in
ModelUpdateService._refresh_single_model() from error (with exc_info)
to warning (message only), since the full traceback is already logged
upstream in CivitaiClient.

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
---
 py/services/civitai_client.py       | 30 +++++++++++++++++++++++++++++
 py/services/model_update_service.py |  3 +--
 2 files changed, 31 insertions(+), 2 deletions(-)

diff --git a/py/services/civitai_client.py b/py/services/civitai_client.py
index 65178cd9..7452f69d 100644
--- a/py/services/civitai_client.py
+++ b/py/services/civitai_client.py
@@ -201,6 +201,29 @@ class CivitaiClient:
 
         return _from_value(payload)
 
+    @staticmethod
+    def _is_transient_server_error(message: str) -> bool:
+        """Return True when the message indicates a transient upstream failure.
+
+        Recognises Cloudflare 524, generic 5xx, and connectivity-level flakiness
+        that should not be treated as a permanent failure.
+        """
+        normalized = message.lower()
+        if "status 5" in normalized or "status 524" in normalized:
+            return True
+        if any(
+            keyword in normalized
+            for keyword in (
+                "connection refused",
+                "connection reset",
+                "temporary failure",
+                "name resolution",
+                "connection closed",
+            )
+        ):
+            return True
+        return False
+
     async def get_model_versions(self, model_id: str) -> Optional[Dict]:
         """Get all versions of a model with local availability info"""
         try:
@@ -223,6 +246,13 @@ class CivitaiClient:
                 logger.info("Civitai request skipped: %s", OFFLINE_FRIENDLY_MESSAGE)
                 return None
             if message:
+                if self._is_transient_server_error(message):
+                    logger.info(
+                        "Transient server error for model %s: %s",
+                        model_id,
+                        message,
+                    )
+                    return None
                 raise RuntimeError(message)
             return None
         except RateLimitError:
diff --git a/py/services/model_update_service.py b/py/services/model_update_service.py
index 0bfcfd23..0965e229 100644
--- a/py/services/model_update_service.py
+++ b/py/services/model_update_service.py
@@ -1000,12 +1000,11 @@ class ModelUpdateService:
                     fallback_error_message = str(exc) or "resource not found"
                     mark_model_as_ignored = True
                 except Exception as exc:  # pragma: no cover - defensive log
-                    logger.error(
+                    logger.warning(
                         "Failed to fetch versions for model %s (%s): %s",
                         model_id,
                         model_type,
                         exc,
-                        exc_info=True,
                     )
                     fallback_error_message = str(exc)
             if response is not None: