Add new setting to diagnostics mode. Minor cleanup.

Author: alfredeen

serve_event_listener/main.py

@@ -125,6 +125,7 @@ def _print_diagnostics(args) -> None:
             "NXDOMAIN_CONFIRMATION_COUNT": getattr(
                 sq, "NXDOMAIN_CONFIRMATION_COUNT", "<n/a>"
             ),
+            "ROLLOUT_GUARD_SECONDS": getattr(sq, "ROLLOUT_GUARD_SECONDS", "<n/a>"),
         }
     except Exception:
         probe_cfg = {
@@ -135,6 +136,7 @@ def _print_diagnostics(args) -> None:
             "DELETED_PROBE_WINDOW": "<unavailable>",
             "DELETED_PROBE_INTERVAL": "<unavailable>",
             "NXDOMAIN_CONFIRMATION_COUNT": "<unavailable>",
+            "ROLLOUT_GUARD_SECONDS": "<unavailable>",
         }
 
     logger.info("\n===== diagnostics =====")

serve_event_listener/probing.py

@@ -59,7 +59,7 @@ def probe_url(self, port80_url: str, headers: dict | None = None) -> ProbeResult
         """Probe a single HTTP URL and classify availability."""
         # NotFound — DNS cannot resolve
         if not self._dns_resolves(port80_url):
-            logger.info(
+            logger.debug(
                 "Skipping URL probing because DNS resolution failed for url %s",
                 port80_url,
             )

serve_event_listener/status_data.py

@@ -335,7 +335,7 @@ def update(self, event: dict) -> None:
             # As we are not ready to introduce a new status Terminating, we will only log it for now
             is_terminating = is_pod_terminating(pod)
             logger.info(
-                "Is release %s identified as Termianting? %s", release, is_terminating
+                "Is release %s identified as Terminating? %s", release, is_terminating
             )
 
             status_object = pod.status

serve_event_listener/status_queue.py

@@ -164,7 +164,7 @@ def process(self) -> None:
             self._rollout_block_until = {}  # release -> epoch (float)
 
         PERIOD = PROCESS_TICK_SECONDS
-        guard_seconds = globals().get("ROLLOUT_GUARD_SECONDS", 15.0)
+        guard_seconds = globals().get("ROLLOUT_GUARD_SECONDS", 30.0)
 
         next_tick = time.monotonic()
         while not self.stop_event.is_set():
@@ -199,6 +199,10 @@ def process(self) -> None:
                     block_until = self._rollout_block_until.get(release)
                     if block_until and now < block_until:
                         # inside guard: defer; optionally probe to see if it flipped back to Running
+                        logger.debug(
+                            "Release %s guarded by rollout block, rejecting preliminary status deleted",
+                            release,
+                        )
                         if (
                             self.prober
                             and self._probe_enabled_for(rec)
@@ -231,11 +235,11 @@ def process(self) -> None:
                         # (no-op here, because we're still inside guard)
                         continue
                     else:
-                        # guard expired → clean out any stale entry
+                        # guard expired, clean out any stale entry
                         if block_until and now >= block_until:
                             self._rollout_block_until.pop(release, None)
 
-                # probing path (gated)
+                # ---- Probing path ----
                 if status_lc in {"running", "deleted"} and self._probe_enabled_for(rec):
 
                     logger.info(
@@ -245,7 +249,7 @@ def process(self) -> None:
                     )
 
                     deadline_epoch = self._ensure_deadline(rec, status_lc)
-                    if deadline_epoch is not None and time.time() < deadline_epoch:
+                    if deadline_epoch is not None and now < deadline_epoch:
                         # still inside the probe window
                         if self._allow_probe_now(rec):
                             # try a probe attempt