queue-runner: reduce the time between queue monitor restarts

This will induce more DB queries (though these are fairly cheap), but at
the benefit of processing bumps within 1m instead of within 10m.

src/hydra-queue-runner/queue-monitor.cc

@@ -318,7 +318,7 @@ bool State::getQueuedBuilds(Connection & conn,
 
         /* Stop after a certain time to allow priority bumps to be
            processed. */
-        if (std::chrono::system_clock::now() > start + std::chrono::seconds(600)) {
+        if (std::chrono::system_clock::now() > start + std::chrono::seconds(60)) {
             prom.queue_checks_early_exits.Increment();
             break;
         }