This side’s produced count over the last REPORT_LAG, one sample per balancer update, so
that a counterparty report can be read net of the production it cannot have seen yet.
Runs a continuous process that attempts to evaluate and
regulate the flow of SURBs to the Session counterparty,
to keep the number of SURBs locally or at the counterparty at a certain level.
Consecutive non-rising reported levels (below target, no distress) before the counterparty’s
store capacity is inferred. Small, but greater than one: a single flat sample is ordinary churn,
and the cost of a false inference is under-supplying the counterparty, so the bar errs high.
How long the refill that follows a wipe of the counterparty’s stock is paced for, at a rate of
one fifth of the target per second (at least one keep-alive’s worth, never above
max_surbs_per_sec; see pacing_cap), so the refill takes about as many seconds.
How recent the last reply must be, relative to the last data packet before the current run of unanswered
data, for a silence to be a reply stream that stopped rather than a Session that gets no replies: one
default notify period.
How long a counterparty’s SURB level report is assumed to have been on its way: the production
of this long before a report arrives is added to the level it carries.