From 2ab6e7843ad7bab92201d07d339872563458590b Mon Sep 17 00:00:00 2001
From: prosolis <5590409+prosolis@users.noreply.github.com>
Date: Mon, 13 Jul 2026 18:19:16 -0700
Subject: [PATCH] roster: a quietly-succeeding ticker looks exactly like a dead
 one
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The push logged nothing on success, so during the first deploy an empty board
and a silent log were indistinguishable from a ticker that never started — and
the debug went looking for a bug that wasn't there. (The snapshot was simply
2 minutes out; the first tick hadn't fired yet.)

Logging every push at INFO would be noise forever. Log the transitions instead:
the first time it works, the moment it breaks, and when it recovers.
---
 internal/plugin/pete_roster.go | 27 ++++++++++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/internal/plugin/pete_roster.go b/internal/plugin/pete_roster.go
index e3761c7..c8900fe 100644
--- a/internal/plugin/pete_roster.go
+++ b/internal/plugin/pete_roster.go
@@ -51,6 +51,14 @@ func (p *AdventurePlugin) peteRosterTicker() {
 	}
 }
 
+// rosterPushOK tracks the last push's outcome so we can log the transitions and
+// nothing else. A push every 2 minutes forever is far too noisy to log at INFO,
+// but total silence is worse: a ticker that is succeeding quietly looks exactly
+// like one that never started, and that ambiguity already cost an operator a
+// wrong-turn debug during the first deploy. So: say something the first time it
+// works, say something when it breaks, say something when it recovers.
+var rosterPushOK bool
+
 func (p *AdventurePlugin) pushRoster() {
 	snap, err := buildRosterSnapshot(time.Now().UTC())
 	if err != nil {
@@ -59,11 +67,24 @@ func (p *AdventurePlugin) pushRoster() {
 	}
 	ctx, cancel := context.WithTimeout(context.Background(), rosterPushTimeout)
 	defer cancel()
+
 	if err := peteclient.PushRoster(ctx, snap); err != nil {
-		// Not an error worth shouting about: the next tick retries by simply
+		// The failure itself is not alarming — the next tick retries by simply
 		// being a fresher snapshot, and if we stay down Pete's board correctly
-		// stops claiming to be live.
-		slog.Warn("roster: push failed, dropping snapshot", "err", err)
+		// stops claiming to be live. Only the *transition* is worth a line.
+		if rosterPushOK {
+			slog.Warn("roster: push failed, board will go stale on Pete", "err", err)
+		} else {
+			slog.Debug("roster: push failed, dropping snapshot", "err", err)
+		}
+		rosterPushOK = false
+		return
+	}
+
+	if !rosterPushOK {
+		slog.Info("roster: board accepted by Pete — live adventurer board is publishing",
+			"adventurers", len(snap.Adventurers))
+		rosterPushOK = true
 	}
 }