Fix misapplication of time.Duration that somehow resulted in continuo…

…us pings

Also remove nprocs from the startup script. A single instance is needed
in order for rate limits to work correctly, and multi-process isn't all
that necessary anyway.

Eventually einhorn will be removed and the unit file will go back to
being a plain socket-activated unit.

distro/linux/relic-einhorn

@@ -29,7 +29,6 @@ case "$1" in
             -d $sock \
             -e /run/relic/einhorn.pid \
             -f /run/relic/einhorn.lock \
-            -n $(nproc) \
             -m manual \
             -q \
             -b 0.0.0.0:6300,r \

server/view_health.go

@@ -49,7 +49,7 @@ func (s *Server) healthCheckLoop() {
 		select {
 		case <-t.C:
 			s.healthCheck()
-			t.Reset(time.Second * time.Duration(interval))
+			t.Reset(interval)
 		case <-s.Closed:
 			break
 		}