Turn one instance's state reply into a single observation row

Pure: no network. The observer never calls provisioning_diff(), and not because the request register does not exist yet: it reads an instance's state whole, with no pairs named, so a comparison here would classify every real pair revocato regardless of how populated the register grows to be. That is a property of what this job reads, not a stage the register will outgrow — version, gate, reachability and expirations are properties of reality alone, and stay available without that comparison.

Usage

observe_instance(server, state, today = Sys.Date())

Arguments

server

The instance name.

state

What module_state() returned: a list with ok, errors, payload and gate.

today

The date expirations are compared against. Injected rather than read from the clock so a fixture cannot turn red on a calendar date for a reason unrelated to the code.

Value

A one-row data frame.

Details

A server that cannot be reached becomes a row rather than an exception, for the same reason provisioning_audit() does it: one instance being down must not hide the state of the others.

scadute counts pairs REDCap should already be refusing. It is the one of the three drifts that needs no desired state, and the boundary is inclusive because REDCap denies on expiration <= TODAY — the day written is already interdicted.