Can cosign retry on transient failure?

[funkypenguin]

Describe the feature

We frequently see connaisseur errors due to transient Docker Hub failures. It would be helpful if we could configure a retry limit for cosign, to minimize disruption from these failures.

Optional: Is your feature request related to a problem? Please describe.

A transient failure (for example, on Docker Hub) will cause connaisseur to deny an image.

Optional: Implementation ideas

Allow the operator to specify a retry interval for the cosign verify process.

Optional: Additional context

Here's an example of a transient failure:

*CONNAISSEUR rejected a request: Unexpected Cosign exception for image "docker.io/myorg/bank-vaults:1.15.2": 
Error: Get "https://auth.docker.io/token?scope=repository%3Amyorg%2Fbank-vaults%3Apull&service=registry.docker.io": 
EOF

[xopham]

@funkypenguin I think such a feature could be added. It appears a bit strange to counteract unavailability of an external resource, but still though does not strike me as unreasonable.

[funkypenguin]

Thank you :) I'm only seeking a retry if the external resource fails on the first attempt, which (sadly) seems not uncommon with Docker Hub...