runner.sh

#!/bin/bash

# Ensure SCRIPT_PATH, LOG_PATH, and URL are defined before or above this block

# --- CONFIGURATION (SET THESE) ---
URL="https://<your_ngrok_url>"   # CHANGE THIS
SCRIPT_PATH="/your/script/path.py"
LOG_PATH="/path/to/your/output/log_file.txt"
# --------------------------------

while true; do
    # Targeted PID check (prevents script from matching/killing itself)
    PID=$(pgrep -f "python3.*$SCRIPT_PATH" | head -n 1)
    
    # 1. Check HTTP status with strict network timeouts
    STATUS=$(curl -o /dev/null -s -w "%{http_code}" --connect-timeout 5 --max-time 10 "$URL/")

    # Trigger restart if process is dead OR server returns non-200
    if [ -z "$PID" ] || [ "$STATUS" != "200" ]; then
        echo "[Monitor] Issue detected (PID: ${PID:-NONE}, HTTP: $STATUS). Restarting script..."

        # Clean Kill Logic
        if [ -n "$PID" ]; then
            kill -9 "$PID" 2>/dev/null
            sleep 1
            if kill -0 "$PID" 2>/dev/null; then
                pkill -9 -f "python3.*$SCRIPT_PATH"
                sleep 1
            fi
        fi

        # Spawn new process and capture its exact PID
        nohup python3 "$SCRIPT_PATH" </dev/null >"$LOG_PATH" 2>&1 &
        NEW_PID=$!
        sleep 2

        # --- RECOVERY & HEALTH CHECK LOOP ---
        echo "[Monitor] Waiting for server to come back online..."
        ATTEMPTS=0
        MAX_ATTEMPTS=15  # Max ~30 seconds of waiting before forcing another restart

        while true; do
            # Safeguard 1: Check if the newly spawned process died immediately on boot
            if ! kill -0 "$NEW_PID" 2>/dev/null; then
                echo "[Monitor] ERROR: script crashed on startup! Retrying..."
                break
            fi

            # Safeguard 2: Check status with network timeouts
            STATUS_CHECK=$(curl -o /dev/null -s -w "%{http_code}" --connect-timeout 5 --max-time 10 "$URL/")
            if [ "$STATUS_CHECK" = "200" ]; then
                echo "[Monitor] Server is back online."
                break  # Recovery successful, return to main monitoring state
            fi

            # Safeguard 3: Break out of inner loop if HTTP stays down too long
            ATTEMPTS=$((ATTEMPTS + 1))
            if [ "$ATTEMPTS" -ge "$MAX_ATTEMPTS" ]; then
                echo "[Monitor] Timeout waiting for 200 OK. Forcing fresh restart cycle..."
                break
            fi

            sleep 2
        done
    fi

    # Monitoring interval for normal operation
    sleep 10
done