rpc: add uptime (backport)

[danielw86dev]

Backported rpc "uptime" from upstream (https://github.com/bitcoin/bitcoin/pull/10400). This is used by various node dashboards.

[chromatic]

Concept ACK and code looks good. How should we document this in doc/rpc-maturity.md? I want to mark it as new but I'm not sure it's UNSTABLE, as it's a backport and the implementation is simple enough to verify.

[patricklodder]

How should we document this in doc/rpc-maturity.md? I want to mark it as new but I'm not sure it's UNSTABLE, as it's a backport and the implementation is simple enough to verify.

Agree. I'd recommend to add it as STABLE as long as we're not worried, and just add a note that this was introduced with 1.15.0.

Reason: I'm not worried about doing any breaking changes to this method. For the others that we added last round as UNSTABLE:

• for getblock we may want to get rid of boolean usage on a major version (perhaps not this one but the next)
• getblockstats was a last minute addition and has open PRs that change output (which will break some popular-but-strict java implementations iirc)
• setmaxconnections has open PRs against it to refactor

So this is not the same.

[chromatic]

That makes sense to me too. @danielw86dev can you add to this PR a change to doc/rpc-maturity.md along the lines of what Patrick proposed? Feel free to add it in a single commit or squash if you prefer--either one would work.