Classify recovered AMQP redeploy errors

scripts/redeploy-fleetd.sh

@@ -116,6 +116,41 @@ check_log_path_matches_plist() {
   ok "log path check: script and plist agree ($resolved_out)"
 }
 
+# Classify ERROR lines in one fresh log region. RabbitMQ's client reports a lost TCP connection
+# through ForgivingExceptionHandler. We only quiet its exact "Connection reset" error after one of
+# fleetd's own recovery listeners later reports a completed AMQP recovery. Every other ERROR, and
+# every unmatched reset, stays unexplained so a failed broker link remains loud.
+classify_amqp_connection_errors() {
+  local log_file="$1" line pending_resets=0
+  REDEPLOY_ERROR_COUNT=0
+  REDEPLOY_RECOVERED_AMQP_ERRORS=0
+  REDEPLOY_UNEXPLAINED_ERRORS=0
+
+  while IFS= read -r line || [ -n "$line" ]; do
+    case "$line" in
+      *' ERROR '*|*' SEVERE '*)
+        REDEPLOY_ERROR_COUNT=$((REDEPLOY_ERROR_COUNT + 1))
+        case "$line" in
+          *'com.rabbitmq.client.impl.ForgivingExceptionHandler'*'Connection reset'*)
+            pending_resets=$((pending_resets + 1))
+            ;;
+          *)
+            REDEPLOY_UNEXPLAINED_ERRORS=$((REDEPLOY_UNEXPLAINED_ERRORS + 1))
+            ;;
+        esac
+        ;;
+      *'AMQP connection recovered; cleared held replies for fresh redelivery'*|*'AMQP lead mailbox connection recovered; cleared held messages for fresh redelivery'*)
+        if [ "$pending_resets" -gt 0 ]; then
+          pending_resets=$((pending_resets - 1))
+          REDEPLOY_RECOVERED_AMQP_ERRORS=$((REDEPLOY_RECOVERED_AMQP_ERRORS + 1))
+        fi
+        ;;
+    esac
+  done < "$log_file"
+
+  REDEPLOY_UNEXPLAINED_ERRORS=$((REDEPLOY_UNEXPLAINED_ERRORS + pending_resets))
+}
+
 # CB-600: sourceable for testing. When this file is SOURCED (not executed) it stops here — nothing
 # below runs — so a test harness can `source` it to call check_log_path_matches_plist (or the
 # other pure helpers above) against a throwaway plist fixture without ever reaching the mutating
@@ -361,15 +396,21 @@ tail -n "+$((RESTART_MARK + 1))" "$OUT" 2>/dev/null \
   | grep -iE 'deferred|classification:|fleet health:|coverage' | tail -8 | sed 's/^/   /' \
   || echo "   (nothing reported)"
 
-# Errors since the restart, anchored to the marker so old noise cannot leak in.
-ERRS="$(tail -n "+$((RESTART_MARK + 1))" "$OUT" 2>/dev/null | grep -cE ' (ERROR|SEVERE) ' || true)"
+# Errors since the restart, anchored to the marker so old noise cannot leak in. Keep the fresh
+# region in a file because the classifier must preserve the order of errors and recoveries.
+FRESH_LOG="$(mktemp -t fleetd-fresh-log)"
+trap 'rm -f "$FRESH_LOG"' EXIT
+tail -n "+$((RESTART_MARK + 1))" "$OUT" > "$FRESH_LOG" 2>/dev/null || true
+classify_amqp_connection_errors "$FRESH_LOG"
 say "result"
 ok "pid $NEW_PID, jar $(jar_id)"
-if [ "${ERRS:-0}" -gt 0 ]; then
-  warn "$ERRS ERROR lines since restart:"
-  tail -n "+$((RESTART_MARK + 1))" "$OUT" | grep -E ' (ERROR|SEVERE) ' | tail -5 | sed 's/^/   /'
-else
+if [ "$REDEPLOY_ERROR_COUNT" -eq 0 ]; then
   ok "no ERROR lines since restart"
+elif [ "$REDEPLOY_UNEXPLAINED_ERRORS" -eq 0 ]; then
+  ok "$REDEPLOY_RECOVERED_AMQP_ERRORS AMQP connection reset ERROR lines recovered since restart"
+else
+  warn "$REDEPLOY_ERROR_COUNT ERROR lines since restart:"
+  grep -E ' (ERROR|SEVERE) ' "$FRESH_LOG" | tail -5 | sed 's/^/   /'
 fi
 echo
 echo "   Next: call fleet_whoami and confirm it still answers 'primary'. A lead whose tab label"

scripts/test-redeploy-fleetd.sh

@@ -0,0 +1,102 @@
+#!/usr/bin/env bash
+# Self-contained checks for the pure log classifier in redeploy-fleetd.sh.
+
+set -euo pipefail
+
+ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+TMP="$(mktemp -d "$ROOT/.redeploy-log-test.XXXXXX")"
+trap 'rm -rf "$TMP"' EXIT
+
+# Sourcing stops before redeploy-fleetd.sh can build, stop, or start the daemon.
+source "$ROOT/scripts/redeploy-fleetd.sh"
+
+fail() {
+  printf 'FAIL: %s\n' "$*" >&2
+  return 1
+}
+
+assert_equals() {
+  local expected="$1" actual="$2" description="$3"
+  [ "$expected" = "$actual" ] || fail "$description: expected $expected, got $actual"
+}
+
+classify_fixture() {
+  local name="$1"
+  classify_amqp_connection_errors "$TMP/$name"
+}
+
+test_no_errors() {
+  cat > "$TMP/no-errors.log" <<'LOG'
+2026-09-05 12:00:00 INFO fleetd listening
+LOG
+  classify_fixture no-errors.log
+  assert_equals 0 "$REDEPLOY_ERROR_COUNT" "no-errors total"
+  assert_equals 0 "$REDEPLOY_UNEXPLAINED_ERRORS" "no-errors unexplained"
+}
+
+test_recovered_connection_error() {
+  cat > "$TMP/recovered.log" <<'LOG'
+2026-09-05 12:00:00 ERROR com.rabbitmq.client.impl.ForgivingExceptionHandler - An unexpected connection driver error occurred (Exception message: Connection reset)
+2026-09-05 12:00:01 INFO dev.ltms.fleet.msg.AmqpReplyInbox - AMQP connection recovered; cleared held replies for fresh redelivery
+LOG
+  classify_fixture recovered.log
+  assert_equals 1 "$REDEPLOY_ERROR_COUNT" "recovered total"
+  assert_equals 1 "$REDEPLOY_RECOVERED_AMQP_ERRORS" "recovered AMQP errors"
+  assert_equals 0 "$REDEPLOY_UNEXPLAINED_ERRORS" "recovered unexplained"
+}
+
+test_unrecovered_connection_error() {
+  cat > "$TMP/unrecovered.log" <<'LOG'
+2026-09-05 12:00:00 ERROR com.rabbitmq.client.impl.ForgivingExceptionHandler - An unexpected connection driver error occurred (Exception message: Connection reset)
+LOG
+  classify_fixture unrecovered.log
+  assert_equals 1 "$REDEPLOY_ERROR_COUNT" "unrecovered total"
+  assert_equals 0 "$REDEPLOY_RECOVERED_AMQP_ERRORS" "unrecovered AMQP errors"
+  assert_equals 1 "$REDEPLOY_UNEXPLAINED_ERRORS" "unrecovered unexplained"
+}
+
+test_other_error_is_unexplained() {
+  cat > "$TMP/other-error.log" <<'LOG'
+2026-09-05 12:00:00 ERROR dev.ltms.fleet.Fleetd - startup failed
+2026-09-05 12:00:01 INFO dev.ltms.fleet.msg.AmqpReplyInbox - AMQP connection recovered; cleared held replies for fresh redelivery
+LOG
+  classify_fixture other-error.log
+  assert_equals 1 "$REDEPLOY_ERROR_COUNT" "other-error total"
+  assert_equals 1 "$REDEPLOY_UNEXPLAINED_ERRORS" "other-error unexplained"
+}
+
+test_mutation_is_caught() {
+  classify_amqp_connection_errors() {
+    local log_file="$1" line
+    REDEPLOY_ERROR_COUNT=0
+    REDEPLOY_RECOVERED_AMQP_ERRORS=0
+    REDEPLOY_UNEXPLAINED_ERRORS=0
+    while IFS= read -r line || [ -n "$line" ]; do
+      case "$line" in
+        *' ERROR '*|*' SEVERE '*)
+          REDEPLOY_ERROR_COUNT=$((REDEPLOY_ERROR_COUNT + 1))
+          case "$line" in
+            *'com.rabbitmq.client.impl.ForgivingExceptionHandler'*'Connection reset'*)
+              REDEPLOY_RECOVERED_AMQP_ERRORS=$((REDEPLOY_RECOVERED_AMQP_ERRORS + 1))
+              ;;
+            *) REDEPLOY_UNEXPLAINED_ERRORS=$((REDEPLOY_UNEXPLAINED_ERRORS + 1)) ;;
+          esac
+          ;;
+      esac
+    done < "$log_file"
+  }
+
+  if test_unrecovered_connection_error > "$TMP/mutation-output" 2>&1; then
+    fail "mutation accepted an unrecovered connection error"
+  fi
+  grep -F 'FAIL: unrecovered AMQP errors: expected 0, got 1' "$TMP/mutation-output" > /dev/null \
+    || fail "mutation failed without the expected assertion"
+  printf 'Mutation check: FAIL: unrecovered AMQP errors: expected 0, got 1\n'
+}
+
+test_no_errors
+test_recovered_connection_error
+test_unrecovered_connection_error
+test_other_error_is_unexplained
+test_mutation_is_caught
+printf 'PASS: redeploy log classifier\n'