Consider have persistent caching DID documents

[reinkrul]

We do have caching for it (mainly for speeding up DID resolving), but only in-memory. If an issuer's DID document becomes unavailable for whatever reason, things break in practice. We could quite easily mitigate this by making the cache persistent (e.g. in SQL or Redis/etc)

[gerardsn]

This is indeed very relevant for issuers
We currently cache the statuslist credentials with the idea that it is available if the issuer has some downtime, but the first step in validating a credential is signature validation that fails if the DID document of the issuer is not available.

[woutslakhorst]

Most policies allow for a 4 hour stale copy of revocationlists/keymaterial which could be bridged by a stale local copy. I would prefer to check if an issuer can run a cluster of nodes first since this would reduce downtime the most.