pub const UNREADY_AFTER: Duration;Expand description
How long a signal has to be unable to store anything before this node calls itself unready.
It has to outlast the automatic recovery, or readiness flaps through every
incident it is supposed to report. The two recoveries are a retention sweep,
which runs every 60s and is what frees a full volume (reclaim), and the
next flush, which is at most max_block_age behind it. Two sweeps gives that
path two chances before the endpoint is pulled — and at a 2s block age it is
already ~60 consecutive failed publishes, which is nobody’s transient.