root
9e87927444
PR Pipeline — Authorize → Validate → Review → Merge / auth (pull_request) Successful in 5s
PR Pipeline — Authorize → Validate → Review → Merge / validate (pull_request) Successful in 5s
PR Pipeline — Authorize → Validate → Review → Merge / lint (pull_request) Successful in 5s
PR Pipeline — Authorize → Validate → Review → Merge / ai-review (pull_request) Successful in 3s
PR Pipeline — Authorize → Validate → Review → Merge / gate (pull_request) Successful in 1s
- Add zulip-watchdog to Maintains (it's running, infrastructure-monitoring expects it) - Remove gpu-monitor from PM2 Maintains (it's systemd-only, not PM2-tracked) - Add Execution steps for all monitored processes (gitea-runner, zulip-watchdog) - Update alert channel: Telegram is primary, Zulip DM is secondary - Update script to check all 4 processes (abiba-telegram, abiba-zulip, gitea-runner, zulip-watchdog) - Add restart count thresholds for all processes - Update log output to include all process statuses
125 lines
5.0 KiB
Bash
Executable File
125 lines
5.0 KiB
Bash
Executable File
#!/bin/bash
|
|
# pm2-self-heal — hourly PM2 process check
|
|
# Part of the pm2-self-heal prose contract
|
|
# Alerts via Telegram (primary) and Zulip DM (secondary, abiba-zulip restored 2026-08-03)
|
|
# Field positions (awk -F'│'): $7=pid $8=uptime $9=restarts $10=status
|
|
|
|
TELEGRAM_BOT_TOKEN="$(grep TELEGRAM_BOT_TOKEN /root/.pi/agent/extensions/telegram/.env 2>/dev/null | cut -d= -f2 || echo '')"
|
|
LOG="/root/pm2-self-heal.log"
|
|
TELEGRAM_CHAT_ID="5822977936"
|
|
|
|
notify_tg() {
|
|
local subject="$1"
|
|
local body="$2"
|
|
local msg="${subject}\n${body}"
|
|
curl -sf -X POST "https://api.telegram.org/bot${TELEGRAM_BOT_TOKEN}/sendMessage" \
|
|
-d "chat_id=${TELEGRAM_CHAT_ID}" \
|
|
-d "text=${msg}" \
|
|
-d "parse_mode=HTML" > /dev/null 2>&1 || true
|
|
}
|
|
|
|
# Log-only mode: replaced by prose contract pm2-self-heal.prose.md
|
|
# Only alerts Telegram on actual failure (status != online)
|
|
# This prevents the process from receiving alerts about itself
|
|
|
|
# Parse PM2 status (--no-color to avoid ANSI escape codes in awk columns)
|
|
STATUS=$(pm2 status --no-color 2>/dev/null)
|
|
|
|
ALERTS=""
|
|
|
|
# Check abiba-telegram (safe to auto-restart)
|
|
TEL_LINE=$(echo "$STATUS" | grep "abiba-telegram")
|
|
TEL_STATUS=$(echo "$TEL_LINE" | awk -F'│' '{print $10}' | xargs)
|
|
TEL_RESTARTS=$(echo "$TEL_LINE" | awk -F'│' '{print $9}' | xargs)
|
|
|
|
if [ "$TEL_STATUS" != "online" ] || [ "$TEL_RESTARTS" -gt 1000 ]; then
|
|
pm2 restart abiba-telegram > /dev/null 2>&1
|
|
sleep 3
|
|
TEL_LINE2=$(pm2 status --no-color 2>/dev/null | grep "abiba-telegram")
|
|
TEL_STATUS2=$(echo "$TEL_LINE2" | awk -F'│' '{print $10}' | xargs)
|
|
if [ "$TEL_STATUS2" = "online" ]; then
|
|
msg="⚠️ abiba-telegram was **$TEL_STATUS** → restarted to online"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
else
|
|
msg="🚨 abiba-telegram **failed restart** (was $TEL_STATUS, still $TEL_STATUS2)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
fi
|
|
|
|
# Check abiba-zulip (live Zulip bridge, heartbeating)
|
|
ZULIP_LINE=$(echo "$STATUS" | grep "abiba-zulip")
|
|
ZULIP_STATUS=$(echo "$ZULIP_LINE" | awk -F'│' '{print $10}' | xargs)
|
|
ZULIP_RESTARTS=$(echo "$ZULIP_LINE" | awk -F'│' '{print $9}' | xargs)
|
|
|
|
if [ "$ZULIP_STATUS" != "online" ]; then
|
|
pm2 restart abiba-zulip > /dev/null 2>&1
|
|
sleep 3
|
|
ZULIP_LINE2=$(pm2 status --no-color 2>/dev/null | grep "abiba-zulip")
|
|
ZULIP_STATUS2=$(echo "$ZULIP_LINE2" | awk -F'│' '{print $10}' | xargs)
|
|
if [ "$ZULIP_STATUS2" = "online" ]; then
|
|
msg="⚠️ abiba-zulip was **$ZULIP_STATUS** → restarted to online"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
else
|
|
msg="🚨 abiba-zulip **failed restart** (was $ZULIP_STATUS, still $ZULIP_STATUS2)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
elif [ "$ZULIP_RESTARTS" -gt 5 ]; then
|
|
msg="⚠️ abiba-zulip has **$ZULIP_RESTARTS** restarts (high count)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
|
|
# Check gitea-runner
|
|
GITEA_LINE=$(echo "$STATUS" | grep "gitea-runner")
|
|
GITEA_STATUS=$(echo "$GITEA_LINE" | awk -F'│' '{print $10}' | xargs)
|
|
GITEA_RESTARTS=$(echo "$GITEA_LINE" | awk -F'│' '{print $9}' | xargs)
|
|
|
|
if [ "$GITEA_STATUS" != "online" ]; then
|
|
pm2 restart gitea-runner > /dev/null 2>&1
|
|
sleep 3
|
|
GITEA_LINE2=$(pm2 status --no-color 2>/dev/null | grep "gitea-runner")
|
|
GITEA_STATUS2=$(echo "$GITEA_LINE2" | awk -F'│' '{print $10}' | xargs)
|
|
if [ "$GITEA_STATUS2" = "online" ]; then
|
|
msg="⚠️ gitea-runner was **$GITEA_STATUS** → restarted to online"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
else
|
|
msg="🚨 gitea-runner **failed restart** (was $GITEA_STATUS, still $GITEA_STATUS2)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
elif [ "$GITEA_RESTARTS" -gt 5 ]; then
|
|
msg="⚠️ gitea-runner has **$GITEA_RESTARTS** restarts (high count)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
|
|
# Check zulip-watchdog
|
|
WATCHDOG_LINE=$(echo "$STATUS" | grep "zulip-watchdog")
|
|
WATCHDOG_STATUS=$(echo "$WATCHDOG_LINE" | awk -F'│' '{print $10}' | xargs)
|
|
WATCHDOG_RESTARTS=$(echo "$WATCHDOG_LINE" | awk -F'│' '{print $9}' | xargs)
|
|
|
|
if [ "$WATCHDOG_STATUS" != "online" ]; then
|
|
pm2 restart zulip-watchdog > /dev/null 2>&1
|
|
sleep 3
|
|
WATCHDOG_LINE2=$(pm2 status --no-color 2>/dev/null | grep "zulip-watchdog")
|
|
WATCHDOG_STATUS2=$(echo "$WATCHDOG_LINE2" | awk -F'│' '{print $10}' | xargs)
|
|
if [ "$WATCHDOG_STATUS2" = "online" ]; then
|
|
msg="⚠️ zulip-watchdog was **$WATCHDOG_STATUS** → restarted to online"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
else
|
|
msg="🚨 zulip-watchdog **failed restart** (was $WATCHDOG_STATUS, still $WATCHDOG_STATUS2)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
elif [ "$WATCHDOG_RESTARTS" -gt 5 ]; then
|
|
msg="⚠️ zulip-watchdog has **$WATCHDOG_RESTARTS** restarts (high count)"
|
|
ALERTS="${ALERTS}${msg}\n"
|
|
fi
|
|
|
|
# Log check
|
|
{
|
|
echo "[$(date '+%Y-%m-%d %H:%M:%S')] tel=$TEL_STATUS zulip=$ZULIP_STATUS gitea=$GITEA_STATUS watchdog=$WATCHDOG_STATUS alerts=${ALERTS:+yes}"
|
|
[ -n "$ALERTS" ] && echo "$ALERTS"
|
|
} >> "$LOG"
|
|
|
|
# Only Telegram DM when there's something to report
|
|
if [ -n "$ALERTS" ]; then
|
|
notify_tg "PM2 Self-Heal — $(date '+%H:%M UTC')" "$ALERTS"
|
|
fi
|