asked again counter

5 files changed, 19 insertions, 3 deletions

diff --git a/docs/measurements.md b/docs/measurements.md
index cff7add..b5c3d6e 100644
--- a/docs/measurements.md
+++ b/docs/measurements.md
@@ -261,6 +261,7 @@ forgotten peer's are gone).
 |---|---|---|
 | `incidents.replyTimeouts` | `peer_reply_timeouts_total` | Our sync messages the peer didn't answer in time (10 s, 30 s with Tor enabled, doubling to 30 times that); resent. Lost messages, or a slow peer. |
 | `incidents.heals` | `peer_heals_total` | Re-handshakes because our heads and the peer's differed for `resync_after` with no sync between ([`peers.md`](peers.md), "Keepalive and self-healing"). Rare in a healthy swarm; frequent means sync state keeps getting out of step. |
+| `incidents.askedAgain` | `peer_asked_again_total` | Times nothing answered what we sent (a keepalive, or a sync message that awaits an answer) for half the time it may before the peer counts as unreached (`unreached_after`, twice the first sync reply timeout), so a PING asked again ([`peers.md`](peers.md)). Each is a PING and its PONG on top of the keepalives. |
 | `incidents.dormancies` | `peer_dormancies_total` | Times the peer went dormant (nothing received for `dormant_after`). |
 | `incidents.dormantMs` | `peer_dormant_seconds_total` | Time it spent dormant, up to its last return. |
 
diff --git a/docs/openapi.yaml b/docs/openapi.yaml
index db73ede..f45fae7 100644
--- a/docs/openapi.yaml
+++ b/docs/openapi.yaml
@@ -1743,7 +1743,7 @@ components:
 
     Incidents:
       type: object
-      required: [replyTimeouts, heals, dormancies, dormantMs]
+      required: [replyTimeouts, heals, askedAgain, dormancies, dormantMs]
       properties:
         replyTimeouts:
           type: integer
@@ -1751,6 +1751,9 @@ components:
         heals:
           type: integer
           description: Re-handshakes because documents differed while sync was idle.
+        askedAgain:
+          type: integer
+          description: Times nothing answered our keepalives or sync messages for a while, so a PING asked again.
         dormancies:
           type: integer
         dormantMs:
diff --git a/src/api/stats/prometheus.rs b/src/api/stats/prometheus.rs
index ffe6593..fb8b890 100644
--- a/src/api/stats/prometheus.rs
+++ b/src/api/stats/prometheus.rs
@@ -1068,7 +1068,7 @@ pub(super) fn render(stats: &StatsResponse, list: &PeerListResponse) -> String {
         }
     }
 
-    let incidents: [IncidentSeries; 4] = [
+    let incidents: [IncidentSeries; 5] = [
         (
             "peer_reply_timeouts_total",
             "Sync messages unanswered until the reply timeout.",
@@ -1080,6 +1080,11 @@ pub(super) fn render(stats: &StatsResponse, list: &PeerListResponse) -> String {
             |i| i.heals as f64,
         ),
         (
+            "peer_asked_again_total",
+            "Times nothing answered for a while, so a PING asked again.",
+            |i| i.asked_again as f64,
+        ),
+        (
             "peer_dormancies_total",
             "Times the peer went dormant.",
             |i| i.dormancies as f64,
diff --git a/src/simtest/bench.rs b/src/simtest/bench.rs
index 06b557e..3b1b3e6 100644
--- a/src/simtest/bench.rs
+++ b/src/simtest/bench.rs
@@ -622,10 +622,12 @@ fn report(
     let (i0, i1) = (during.0.incidents, during.1.incidents);
     let _ = writeln!(
         out,
-        "    incidents: {} reply timeouts ({:.2} per write), {} heals, {} dormancies",
+        "    incidents: {} reply timeouts ({:.2} per write), {} heals, {} asked again ({:.1} per write), {} dormancies",
         i1.reply_timeouts - i0.reply_timeouts,
         (i1.reply_timeouts - i0.reply_timeouts) as f64 / writes,
         i1.heals - i0.heals,
+        i1.asked_again - i0.asked_again,
+        (i1.asked_again - i0.asked_again) as f64 / writes,
         i1.dormancies - i0.dormancies,
     );
     let mut propagation = Propagation::default();
diff --git a/src/state/peer.rs b/src/state/peer.rs
index 787a772..a84e069 100644
--- a/src/state/peer.rs
+++ b/src/state/peer.rs
@@ -233,6 +233,9 @@ pub(crate) struct Incidents {
     pub(crate) reply_timeouts: u64,
     /// Re-handshakes because our documents differed while sync was idle.
     pub(crate) heals: u64,
+    /// Times nothing answered us for half the time it may (`AwaitingAnswer`), so we
+    /// asked again, with a PING (or the keepalive then due).
+    pub(crate) asked_again: u64,
     /// Times the peer went dormant.
     pub(crate) dormancies: u64,
     /// Time spent dormant, in milliseconds, up to its last return (the current
@@ -244,6 +247,7 @@ impl Incidents {
     pub(crate) fn add(&mut self, other: &Incidents) {
         self.reply_timeouts += other.reply_timeouts;
         self.heals += other.heals;
+        self.asked_again += other.asked_again;
         self.dormancies += other.dormancies;
         self.dormant_ms += other.dormant_ms;
     }
@@ -439,6 +443,7 @@ impl PeerState {
         match &mut self.awaiting_answer {
             Some(awaiting) if awaiting.ask_again_at.is_some_and(|at| now >= at) => {
                 awaiting.ask_again_at = None;
+                self.incidents.asked_again += 1;
                 true
             }
             _ => false,