I would try to find an alternative first. Google Scholar often lists three useful links to entries:
Cited by ...,
Related articles and
All ... versions. Checking them out often resolves the problem: you either get access to the material, or find an equal (or sometimes better) alternative.
Similarly, Crossref allows you to filter on the publication (
Actions menu) helping to find an alternative.
In general, I also agree that reading the abstract and/or supplementary information is already a good reason to cite the source. However, I would strongly discourage to cite that web page unless you are absolutely sure it won't be gone in a matter of months. Maybe refer to the unique identifier such as DOI, or think about adding a WayBackMachine's URL of that site. Either way, it's safe to add a date when the webpage has been accessed (in the bibliography entry).