Director stops synchronizing with registry if there is a problem

[mguidon]

Problem Description

I published services into osparc.speag.com but they never showed up. Upon inspection I realized that for a unknown reason, one image was missing in the registry since a few weeks (unrelated to the new things I wanted to push). The director notices that and throws an exception but does not continue its job. In the log I could see the below error messages repeated every few seconds:

2024-09-06T12:56:47.091075774Z NoneType: None
2024-09-06T12:56:47.091596657Z ERROR:simcore_service_director.rest.generated_code.routing-38: error happened in handling route
2024-09-06T12:56:47.091613469Z Traceback (most recent call last):
2024-09-06T12:56:47.091619823Z   File "/home/scu/.venv/lib/python3.6/site-packages/simcore_service_director/rest/generated_code/routing.py", line 29, in __handle_errors
2024-09-06T12:56:47.091625937Z     response = await handler(request)
2024-09-06T12:56:47.091631099Z   File "/home/scu/.venv/lib/python3.6/site-packages/aiohttp_apiset/swagger/route.py", line 90, in handler
2024-09-06T12:56:47.091637985Z     response = await self._handler(**parameters)
2024-09-06T12:56:47.091646875Z   File "/home/scu/.venv/lib/python3.6/site-packages/simcore_service_director/rest/handlers.py", line 130, in service_extras_by_key_version_get
2024-09-06T12:56:47.091653711Z     raise web_exceptions.HTTPNotFound(reason=str(err))
2024-09-06T12:56:47.091657714Z aiohttp.web_exceptions.HTTPNotFound: The service /v2/simcore/services/dynamic/s4l-core-8-0-0-modeling/manifests/3.2.34:UNDEFINED does not exist
2024-09-06T12:56:49.299534775Z ERROR:simcore_service_director.registry_proxy-73: Path to registry not found: https://registry.osparc.speag.com/v2/simcore/services/dynamic/s4l-core-8-0-0-modeling/manifests/3.2.34
2024-09-06T12:56:49.299543124Z NoneType: None
2024-09-06T12:56:49.300012016Z ERROR:simcore_service_director.rest.generated_code.routing-38: error happened in handling route
2024-09-06T12:56:49.300017004Z Traceback (most recent call last):
2024-09-06T12:56:49.300019613Z   File "/home/scu/.venv/lib/python3.6/site-packages/simcore_service_director/rest/generated_code/routing.py", line 29, in __handle_errors
2024-09-06T12:56:49.300022426Z     response = await handler(request)
2024-09-06T12:56:49.300025046Z   File "/home/scu/.venv/lib/python3.6/site-packages/aiohttp_apiset/swagger/route.py", line 90, in handler
2024-09-06T12:56:49.300027020Z     response = await self._handler(**parameters)
2024-09-06T12:56:49.300028941Z   File "/home/scu/.venv/lib/python3.6/site-packages/simcore_service_director/rest/handlers.py", line 130, in service_extras_by_key_version_get
2024-09-06T12:56:49.300031504Z     raise web_exceptions.HTTPNotFound(reason=str(err))
2024-09-06T12:56:49.300033588Z aiohttp.web_exceptions.HTTPNotFound: The service /v2/simcore/services/dynamic/s4l-core-8-0-0-modeling/manifests/3.2.34:UNDEFINED does not exist

After I manually pushed the missing image, the director became happy again and the catalog started listing the services.

I am aware of the fact that this is first and foremost a curation issue for service publication but I still believe that one such error should not bring down this whole part of the platform.