From fd57e75cbf3c88189caaa9f4a1a074fd67f07433 Mon Sep 17 00:00:00 2001
From: Bug1 Repro <skylerah@noreply.embr>
Date: Fri, 29 May 2026 10:22:53 -0700
Subject: [PATCH 1/2] Bug #1 repro: crash-loop start to test supervisor port
 drain

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 build.yaml       |  3 ++-
 flaky-startup.sh | 27 +++++++++++++++++++++++++++
 2 files changed, 29 insertions(+), 1 deletion(-)
 create mode 100755 flaky-startup.sh

diff --git a/build.yaml b/build.yaml
index 4aea474..52447b5 100644
--- a/build.yaml
+++ b/build.yaml
@@ -2,4 +2,5 @@ version: 1
 platform: python
 platformVersion: "3.14"
 run:
-  port: 8080
\ No newline at end of file
+  port: 8080
+  startCommand: "bash /output/flaky-startup.sh"
diff --git a/flaky-startup.sh b/flaky-startup.sh
new file mode 100755
index 0000000..0f1ab1f
--- /dev/null
+++ b/flaky-startup.sh
@@ -0,0 +1,27 @@
+#!/bin/bash
+set +e
+TAG="${EMBR_DEPLOYMENT_ID:-unknown}"
+echo "[flaky-startup ${TAG}] starting at $(date -u)"
+echo "[flaky-startup ${TAG}] PID=$$, PPID=$PPID"
+echo "[flaky-startup ${TAG}] ports in use BEFORE bind:"
+ss -ltnp 2>&1 | grep -E ':8080|LISTEN' | head -10
+
+python -c "
+import socket, sys, time, os
+s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+try:
+    s.bind(('0.0.0.0', 8080))
+    s.listen(8)
+    print(f'[flaky-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] BOUND :8080 successfully', flush=True)
+except OSError as e:
+    print(f'[flaky-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] BIND FAILED: {e}', flush=True)
+    sys.exit(98)
+time.sleep(12)
+print(f'[flaky-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] exiting non-zero now', flush=True)
+sys.exit(1)
+"
+EXITCODE=$?
+echo "[flaky-startup ${TAG}] python exited ${EXITCODE} at $(date -u); supervisor will restart"
+sleep 1
+exit $EXITCODE

From e0f22e4709155dc6e3110893429585c43a200220 Mon Sep 17 00:00:00 2001
From: Bug1 Repro <skylerah@noreply.embr>
Date: Fri, 29 May 2026 10:27:12 -0700
Subject: [PATCH 2/2] Bug #1 repro rev 2: healthy startup with bind-retry
 instrumentation

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 flaky-startup.sh | 52 ++++++++++++++++++++++++++++++------------------
 1 file changed, 33 insertions(+), 19 deletions(-)

diff --git a/flaky-startup.sh b/flaky-startup.sh
index 0f1ab1f..590ee0e 100755
--- a/flaky-startup.sh
+++ b/flaky-startup.sh
@@ -1,27 +1,41 @@
 #!/bin/bash
 set +e
 TAG="${EMBR_DEPLOYMENT_ID:-unknown}"
-echo "[flaky-startup ${TAG}] starting at $(date -u)"
-echo "[flaky-startup ${TAG}] PID=$$, PPID=$PPID"
-echo "[flaky-startup ${TAG}] ports in use BEFORE bind:"
+echo "[healthy-startup ${TAG}] starting at $(date -u)"
+echo "[healthy-startup ${TAG}] PID=$$, PPID=$PPID"
+echo "[healthy-startup ${TAG}] ports in use BEFORE bind:"
 ss -ltnp 2>&1 | grep -E ':8080|LISTEN' | head -10
+echo "[healthy-startup ${TAG}] my parent process tree:"
+ps -ef --forest 2>&1 | head -30
 
 python -c "
 import socket, sys, time, os
-s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
-s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
-try:
-    s.bind(('0.0.0.0', 8080))
-    s.listen(8)
-    print(f'[flaky-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] BOUND :8080 successfully', flush=True)
-except OSError as e:
-    print(f'[flaky-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] BIND FAILED: {e}', flush=True)
+print(f'[healthy-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] attempting bind on :8080', flush=True)
+attempts = 0
+while attempts < 30:
+    s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+    try:
+        s.bind(('0.0.0.0', 8080))
+        s.listen(8)
+        print(f'[healthy-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] BOUND :8080 successfully after {attempts} attempts', flush=True)
+        break
+    except OSError as e:
+        attempts += 1
+        print(f'[healthy-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] BIND FAILED attempt {attempts}: {e}', flush=True)
+        s.close()
+        time.sleep(1)
+else:
+    print(f'[healthy-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] GAVE UP after 30 attempts', flush=True)
     sys.exit(98)
-time.sleep(12)
-print(f'[flaky-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] exiting non-zero now', flush=True)
-sys.exit(1)
-"
-EXITCODE=$?
-echo "[flaky-startup ${TAG}] python exited ${EXITCODE} at $(date -u); supervisor will restart"
-sleep 1
-exit $EXITCODE
+print(f'[healthy-startup {os.environ.get(\"EMBR_DEPLOYMENT_ID\",\"?\")}] serving forever now', flush=True)
+# minimal HTTP responder
+from http.server import HTTPServer, BaseHTTPRequestHandler
+class H(BaseHTTPRequestHandler):
+    def do_GET(self):
+        self.send_response(200); self.end_headers(); self.wfile.write(b'ok')
+    def log_message(self, *a): pass
+HTTPServer.allow_reuse_address = True
+HTTPServer(('0.0.0.0', 8080), H, bind_and_activate=False).serve_forever()
+" || true
+sleep 60