Debug production titiler-cmr deployment

[hrodmn]

I noticed that we are getting some unexpected errors on the new production deployment of titiler-cmr. For example, this request to the staging endpoint returns successfully:
https://staging.openveda.cloud/api/titiler-cmr/compatibility?concept_id=C2036881735-POCLOUD

But the same request to the production endpoint returns 500:

https://openveda.cloud/api/titiler-cmr/compatibility?concept_id=C2036881735-POCLOUD

I tried to find some more detailed logging from the request to the production endpoint (in MCP VEDA AWS account) but there are not any CloudWatch logs being written for the titiler-cmr function! @anayeaye pointed out that this is not a problem for any of the other titiler functions so there might be some interaction between the titiler-cmr infrastructure code and the MCP environment. I know that @ividito had to implement some changes to get the function deployed in the MCP account at all so I wonder if there might be some more MCP-specific changes we need to make to get CloudWatch hooked up properly.