Harden Copilot Setup proxy readiness to prevent false CI failures

[Copilot]

The Copilot Setup Steps workflow was failing before Gradle warmup because the local Google Maven proxy readiness check timed out too aggressively. This change makes proxy startup tolerant to slow initialization and surfaces actionable diagnostics when startup actually fails.

• Root cause addressed

  • Proxy health check previously allowed only ~30s (seq 1 30), which was insufficient in slower runner/network conditions, causing curl to fail and the job to exit early.

• Workflow hardening

  • Capture proxy PID immediately after launch.
  • Increase readiness window to 180 seconds.
  • Keep polling health endpoint until ready while verifying the proxy process is still alive.
  • Fail with explicit log output if the proxy exits early or never becomes ready.

• Failure diagnostics improved

  • Use curl --show-error --fail for clearer failure output.
  • Print /tmp/androidsa-google-maven-proxy.log on readiness failure paths to expose startup errors directly in job logs.

nohup python3 tools/google_maven_proxy.py --host 127.0.0.1 --port 38473 >/tmp/androidsa-google-maven-proxy.log 2>&1 &
proxy_pid="$!"

for attempt in $(seq 1 180); do
  if curl --silent --show-error --fail http://127.0.0.1:38473/ >/dev/null; then
    break
  fi
  if ! kill -0 "$proxy_pid" 2>/dev/null; then
    echo "Local Google Maven proxy exited before becoming ready."
    cat /tmp/androidsa-google-maven-proxy.log
    exit 1
  fi
  sleep 1
done