implement only monitoring dns blocklists, without using them for incoming deliveries

so you can still know when someone has put you on their blocklist (which may
affect delivery), without using them.

also query dnsbls for our ips more often when we do more outgoing connections
for delivery: once every 100 messages, but at least 5 mins and at most 3 hours
since the previous check.

queue/direct.go

@@ -10,6 +10,7 @@ import (
 	"net"
 	"os"
 	"strings"
+	"sync/atomic"
 	"time"
 
 	"github.com/prometheus/client_golang/prometheus"
@@ -30,6 +31,10 @@ import (
 	"github.com/mjl-/mox/tlsrpt"
 )
 
+// Increased each time an outgoing connection is made for direct delivery. Used by
+// dnsbl monitoring to pace querying.
+var connectionCounter atomic.Int64
+
 var (
 	metricDestinations = promauto.NewCounter(
 		prometheus.CounterOpts{
@@ -88,6 +93,10 @@ var (
 	)
 )
 
+func ConnectionCounter() int64 {
+	return connectionCounter.Load()
+}
+
 // todo: rename function, perhaps put some of the params in a delivery struct so we don't pass all the params all the time?
 func fail(ctx context.Context, qlog mlog.Log, m Msg, backoff time.Duration, permanent bool, remoteMTA dsn.NameIP, secodeOpt, errmsg, firstLine string, moreLines []string) {
 	// todo future: when we implement relaying, we should be able to send DSNs to non-local users. and possibly specify a null mailfrom. ../rfc/5321:1503
@@ -534,6 +543,7 @@ func deliverHost(log mlog.Log, resolver dns.Resolver, dialer smtpclient.Dialer,
 		if m.DialedIPs == nil {
 			m.DialedIPs = map[string][]net.IP{}
 		}
+		connectionCounter.Add(1)
 		conn, remoteIP, err = smtpclient.Dial(ctx, log.Logger, dialer, host, ips, 25, m.DialedIPs, mox.Conf.Static.SpecifiedSMTPListenIPs)
 	}
 	cancel()