#!/bin/bash
# Watchdog: called by crontab every minute to ensure the monitor script is alive

MONITOR_SCRIPT="/home/jenkins/.npu-scripts/npu_auto_occupy.sh"
PID_FILE="/home/jenkins/.npu-scripts/npu_auto_occupy.pid"

# Colors
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[0;33m'
NC='\033[0m'

is_running() {
    if [ -f "$PID_FILE" ]; then
        local pid=$(cat "$PID_FILE")
        if kill -0 "$pid" 2>/dev/null && grep -q "npu_auto_occupy.sh" /proc/"$pid"/cmdline 2>/dev/null; then
            return 0
        else
            rm -f "$PID_FILE"
        fi
    fi
    pgrep -f "npu_auto_occupy.sh" > /dev/null 2>&1
}

if ! is_running; then
    # Start monitor script, redirect output to log
    echo -e "${YELLOW}[$(date)] Watchdog: Monitor script not running, starting it...${NC}" >> /home/jenkins/.npu-scripts/npu_auto_occupy.log
    nohup "$MONITOR_SCRIPT" >> /home/jenkins/.npu-scripts/npu_auto_occupy.log 2>&1 &
    # Wait for PID file to be generated (max 5 seconds)
    for i in {1..5}; do
        if [ -f "$PID_FILE" ]; then
            echo -e "${GREEN}[$(date)] Watchdog: Monitor script started, PID file created${NC}" >> /home/jenkins/.npu-scripts/npu_auto_occupy.log
            break
        fi
        sleep 1
    done
fi