Regularly check availability of linked datasets in catalog

[jbusecke]

With a growing number of datasets in our catalog I believe we need a way to regularly check that the datasets are healthy (and potentially an automated way to notify the maintainers if not).

I wonder if we could as a first step just loop over every dataset, try to execute the snippet we provide and list urls that are problematic?

As a follow up step it would be really useful if this CI feature could automatically raise issues on the corresponding feedstock repos, alerting the dataset experts over there (who might not follow this repo as closely).

[andersy005]

We're currently running some of these checks.. well, kind of

this is how i discovered the most recent OSN failure in #160

i think the missing piece is surfacing failures from the CI into an issue that can be tracked...

[jbusecke]

i think the missing piece is surfacing failures from the CI into an issue that can be tracked...

If this could automatically raise an issue and tag @leap-stc/data-and-compute, that would be ideal!