1// monitor.js — collecteur Écosystème Groupe KA2// Health checks HTTPS (45 s) + certificats TLS, sweep des nœuds (5 min :3// charge, RAM, disque, PM2, dernier commit), incidents, historique SQLite,4// push temps réel via le WebSocket du backend.5import https from 'node:https';6import path from 'node:path';7import { execFile } from 'node:child_process';8import { DatabaseSync } from 'node:sqlite';910// Sites du Groupe KA (source : cluster-deployments.json 2026-08-19)11export const SITES = [12 { app: 'lou-ka', label: 'Lou·Ka', domain: 'www.lou-ka.com', node: 'M3U96b', pm2: 'lou-ka-web', dir: '~/apps/lou-ka', deployed: '2026-08' },13 { app: 'immo-ka', label: 'Immo·Ka', domain: 'www.immo-ka.com', node: 'M4M64a', pm2: 'immo-ka-web', dir: '~/apps/immo-ka', deployed: '2026-08' },14 { app: 'vrai-prix', label: 'Vrai-Prix', domain: 'www.vrai-prix.com', node: 'M3U96a', pm2: 'vrai-prix', dir: '~/apps/vrai-prix', deployed: '2026-08' },15 { app: 'valoplex', label: 'ValoPlex', domain: 'www.valoplex.com', node: 'M3U96a', pm2: 'valoplex', dir: '~/apps/valoplex', deployed: '2026-08' },16 { app: 'auto-ka', label: 'Auto·Ka', domain: 'www.auto-ka.com', node: 'M4M64b', pm2: 'auto-ka-web', dir: '~/auto-ka', deployed: '2026-08' },17 { app: 'fabri-ka', label: 'Fabri·Ka', domain: 'www.fabri-ka.com', node: 'M4M64a', pm2: 'fabri-ka-web', dir: '~/fabri-ka', deployed: '2026-08' },18 { app: 'food-ka', label: 'Food·Ka', domain: 'www.food-ka.com', node: 'M4M64b', pm2: 'food-ka-web', dir: '~/apps/food-ka', deployed: '2026-08' },19 { app: 'resto-ka', label: 'Resto·Ka', domain: 'www.resto-ka.com', node: 'M3U96b', pm2: 'resto-ka', dir: '~/apps/resto-ka', deployed: '2026-08' },20 { app: 'sorti-ka', label: 'Sorti·Ka', domain: 'www.sorti-ka.com', node: 'M3U96a', pm2: 'sorti-ka-web', dir: '~/apps/sorti-ka', deployed: '2026-08' },21 { app: 'crea-ka', label: 'Créa·Ka', domain: 'www.crea-ka.com', node: 'M3U96b', pm2: 'crea-ka-web', dir: '~/apps/crea-ka', deployed: '2026-08' },22 { app: 'job-ka', label: 'Job·Ka', domain: 'www.job-ka.com', node: 'M3U96a', pm2: 'job-ka-web', dir: '~/apps/job-ka', deployed: '2026-08' },23 { app: 'trouve-ka', label: 'Trouve·Ka', domain: 'www.trouve-ka.com', node: 'M2M32', pm2: 'tk-web', dir: '~/trouve-ka', deployed: '2026-08' },24 { app: 'groupe-ka', label: 'Groupe·Ka', domain: 'www.groupe-ka.com', node: 'M3U96b', pm2: 'groupe-ka', dir: '~/apps/groupe-ka', deployed: '2026-08' },25];2627const NODES = ['M3U96a', 'M3U96b', 'M4M64a', 'M4M64b', 'M2M32'];28const SELF_NODE = 'M3U96a';29const CHECK_MS = 45_000;30const SWEEP_MS = 5 * 60_000;31const SLOW_MS = 4000; // seuil "lent" (▲)32const FAILS_FOR_INCIDENT = 2; // échecs consécutifs avant incident3334let db, emit;35const consecFails = {};3637export function startMonitor({ dataDir, onEvent }) {38 emit = onEvent || (() => {});39 db = new DatabaseSync(path.join(dataDir, 'eco.sqlite'));40 db.exec(`41 CREATE TABLE IF NOT EXISTS checks(site TEXT, ts INTEGER, ms INTEGER, code INTEGER, ok INTEGER, err TEXT);42 CREATE INDEX IF NOT EXISTS idx_checks ON checks(site, ts);43 CREATE TABLE IF NOT EXISTS incidents(id INTEGER PRIMARY KEY AUTOINCREMENT, site TEXT, started INTEGER, ended INTEGER, reason TEXT);44 CREATE TABLE IF NOT EXISTS cert(site TEXT PRIMARY KEY, expires INTEGER, issuer TEXT, checked INTEGER);45 CREATE TABLE IF NOT EXISTS nodestats(node TEXT, ts INTEGER, load1 REAL, mem_used INTEGER, mem_total INTEGER, disk_pct INTEGER);46 CREATE INDEX IF NOT EXISTS idx_nodestats ON nodestats(node, ts);47 CREATE TABLE IF NOT EXISTS appstats(app TEXT, ts INTEGER, cpu REAL, mem INTEGER, status TEXT, restarts INTEGER, last_commit INTEGER);48 CREATE INDEX IF NOT EXISTS idx_appstats ON appstats(app, ts);49 `);50 checkAll();51 sweepNodes();52 setInterval(checkAll, CHECK_MS);53 setInterval(sweepNodes, SWEEP_MS);54 setInterval(cleanup, 6 * 3600_000);55}5657// ---------- health checks ----------58function checkOne(site) {59 return new Promise((resolve) => {60 const t0 = Date.now();61 const req = https.get({ host: site.domain, path: '/', timeout: 12_000, headers: { 'User-Agent': 'admin-ka-monitor/2.0' } }, (res) => {62 const ms = Date.now() - t0;63 const cert = res.socket.getPeerCertificate?.();64 if (cert && cert.valid_to) {65 db.prepare('INSERT INTO cert(site,expires,issuer,checked) VALUES(?,?,?,?) ON CONFLICT(site) DO UPDATE SET expires=excluded.expires,issuer=excluded.issuer,checked=excluded.checked')66 .run(site.app, Date.parse(cert.valid_to), cert.issuer?.O || '', Date.now());67 }68 res.resume();69 resolve({ site: site.app, ts: Date.now(), ms, code: res.statusCode, ok: res.statusCode >= 200 && res.statusCode < 400 ? 1 : 0, err: null });70 });71 req.on('timeout', () => { req.destroy(new Error('timeout')); });72 req.on('error', (e) => resolve({ site: site.app, ts: Date.now(), ms: Date.now() - t0, code: 0, ok: 0, err: e.message.slice(0, 120) }));73 });74}7576async function checkAll() {77 const results = await Promise.all(SITES.map(checkOne));78 const ins = db.prepare('INSERT INTO checks(site,ts,ms,code,ok,err) VALUES(?,?,?,?,?,?)');79 for (const r of results) {80 ins.run(r.site, r.ts, r.ms, r.code, r.ok, r.err);81 handleIncident(r);82 }83 emit({ type: 'eco', kind: 'checks', items: results });84}8586function handleIncident(r) {87 const open = db.prepare('SELECT id FROM incidents WHERE site=? AND ended IS NULL').get(r.site);88 if (!r.ok) {89 consecFails[r.site] = (consecFails[r.site] || 0) + 1;90 if (consecFails[r.site] >= FAILS_FOR_INCIDENT && !open) {91 db.prepare('INSERT INTO incidents(site,started,reason) VALUES(?,?,?)')92 .run(r.site, r.ts, r.err || `HTTP ${r.code}`);93 emit({ type: 'eco_alert', kind: 'down', site: r.site, reason: r.err || `HTTP ${r.code}`, ts: r.ts });94 }95 } else {96 consecFails[r.site] = 0;97 if (open) {98 db.prepare('UPDATE incidents SET ended=? WHERE id=?').run(r.ts, open.id);99 emit({ type: 'eco_alert', kind: 'up', site: r.site, ts: r.ts });100 }101 }102}103104// ---------- sweep des nœuds (charge, RAM, disque, PM2, dernier commit) ----------105function sweepNode(node) {106 return new Promise((resolve) => {107 const gitCmds = SITES.filter(s => s.node === node)108 .map(s => `echo "GIT ${s.app} $(git -C ${s.dir} log -1 --format=%ct 2>/dev/null || echo 0)"`).join('; ');109 const script = `export PATH=/opt/homebrew/bin:/usr/local/bin:$PATH110echo "LOAD $(sysctl -n vm.loadavg 2>/dev/null)"111echo "MEMTOTAL $(sysctl -n hw.memsize 2>/dev/null)"112echo "PAGESIZE $(sysctl -n hw.pagesize 2>/dev/null)"113vm_stat 2>/dev/null | egrep "active|wired|occupied" | tr -d '.'114echo "DISK $(df -k / | tail -1 | awk '{print $5}' | tr -d '%')"115echo "PM2JSON $(pm2 jlist 2>/dev/null | tail -1)"116${gitCmds}`;117 const args = node === SELF_NODE118 ? ['-c', script]119 : ['-o', 'BatchMode=yes', '-o', 'ConnectTimeout=12', node, script];120 const bin = node === SELF_NODE ? '/bin/bash' : 'ssh';121 execFile(bin, args, { timeout: 45_000, maxBuffer: 8 * 1024 * 1024, env: { ...process.env, PATH: '/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin' } }, (err, stdout) => {122 if (err && !stdout) return resolve(null);123 try { parseSweep(node, stdout); } catch {}124 resolve(null);125 });126 });127}128129function parseSweep(node, out) {130 const ts = Date.now();131 const load = out.match(/^LOAD \{?\s*([\d.]+)/m);132 const memTotal = Number(out.match(/^MEMTOTAL (\d+)/m)?.[1] || 0);133 const pageSize = Number(out.match(/^PAGESIZE (\d+)/m)?.[1] || 16384);134 let pages = 0;135 for (const m of out.matchAll(/Pages (?:active|wired down|occupied by compressor):\s+(\d+)/g)) pages += Number(m[1]);136 const memUsed = pages * pageSize;137 const disk = Number(out.match(/^DISK (\d+)/m)?.[1] || 0);138 db.prepare('INSERT INTO nodestats(node,ts,load1,mem_used,mem_total,disk_pct) VALUES(?,?,?,?,?,?)')139 .run(node, ts, load ? Number(load[1]) : 0, memUsed, memTotal, disk);140141 const commits = {};142 for (const m of out.matchAll(/^GIT (\S+) (\d+)/gm)) commits[m[1]] = Number(m[2]) * 1000;143144 const pm2raw = out.match(/^PM2JSON (\[.*\])\s*$/m);145 const ins = db.prepare('INSERT INTO appstats(app,ts,cpu,mem,status,restarts,last_commit) VALUES(?,?,?,?,?,?,?)');146 if (pm2raw) {147 let list = [];148 try { list = JSON.parse(pm2raw[1]); } catch {}149 for (const site of SITES.filter(s => s.node === node)) {150 const p = list.find(x => x.name === site.pm2) || list.find(x => (x.name || '').startsWith(site.app));151 ins.run(site.app, ts,152 p ? (p.monit?.cpu ?? 0) : 0,153 p ? (p.monit?.memory ?? 0) : 0,154 p ? (p.pm2_env?.status || 'inconnu') : 'introuvable',155 p ? (p.pm2_env?.restart_time ?? 0) : 0,156 commits[site.app] || 0);157 }158 } else {159 for (const site of SITES.filter(s => s.node === node)) {160 ins.run(site.app, ts, 0, 0, 'inconnu', 0, commits[site.app] || 0);161 }162 }163}164165async function sweepNodes() {166 await Promise.all(NODES.map(sweepNode));167 emit({ type: 'eco', kind: 'sweep', ts: Date.now() });168}169170function cleanup() {171 const cut = Date.now() - 35 * 24 * 3600_000;172 db.prepare('DELETE FROM checks WHERE ts < ?').run(cut);173 db.prepare('DELETE FROM nodestats WHERE ts < ?').run(cut);174 db.prepare('DELETE FROM appstats WHERE ts < ?').run(cut);175}176177// ---------- requêtes ----------178function uptime(site, sinceMs) {179 const r = db.prepare('SELECT COUNT(*) n, SUM(ok) up FROM checks WHERE site=? AND ts>?').get(site, Date.now() - sinceMs);180 return r.n ? Math.round((r.up / r.n) * 1000) / 10 : null;181}182function p95(site, sinceMs) {183 const rows = db.prepare('SELECT ms FROM checks WHERE site=? AND ts>? AND ok=1 ORDER BY ms').all(site, Date.now() - sinceMs);184 if (!rows.length) return null;185 return rows[Math.min(rows.length - 1, Math.floor(rows.length * 0.95))].ms;186}187188export function ecoSummary() {189 const out = [];190 for (const s of SITES) {191 const last = db.prepare('SELECT ts,ms,code,ok,err FROM checks WHERE site=? ORDER BY ts DESC LIMIT 1').get(s.app);192 const spark = db.prepare('SELECT ts,ms,ok FROM checks WHERE site=? ORDER BY ts DESC LIMIT 40').all(s.app).reverse();193 const avg = db.prepare('SELECT AVG(ms) a FROM checks WHERE site=? AND ok=1 AND ts>?').get(s.app, Date.now() - 24 * 3600_000);194 const cert = db.prepare('SELECT expires FROM cert WHERE site=?').get(s.app);195 const inc = db.prepare('SELECT started,reason FROM incidents WHERE site=? AND ended IS NULL').get(s.app);196 const stat = db.prepare('SELECT cpu,mem,status,restarts,last_commit,ts FROM appstats WHERE app=? ORDER BY ts DESC LIMIT 1').get(s.app);197 const nodeStat = db.prepare('SELECT load1,mem_used,mem_total,disk_pct,ts FROM nodestats WHERE node=? ORDER BY ts DESC LIMIT 1').get(s.node);198 out.push({199 app: s.app, label: s.label, domain: s.domain, node: s.node,200 last, spark,201 uptime24: uptime(s.app, 24 * 3600_000),202 uptime7: uptime(s.app, 7 * 24 * 3600_000),203 uptime30: uptime(s.app, 30 * 24 * 3600_000),204 avg24: avg?.a ? Math.round(avg.a) : null,205 p95_24: p95(s.app, 24 * 3600_000),206 certExpires: cert?.expires || null,207 incident: inc || null,208 proc: stat || null,209 nodeStat: nodeStat || null,210 });211 }212 return { sites: out, ts: Date.now() };213}214215export function ecoSite(app) {216 const s = SITES.find(x => x.app === app);217 if (!s) return null;218 const since = Date.now() - 24 * 3600_000;219 // ~200 seaux sur 24 h pour le graphique220 const bucket = Math.ceil((24 * 3600_000) / 200);221 const series = db.prepare(`222 SELECT (ts/${bucket})*${bucket} AS t, AVG(ms) ms, MIN(ok) ok, COUNT(*) n223 FROM checks WHERE site=? AND ts>? GROUP BY t ORDER BY t`).all(app, since);224 const incidents = db.prepare('SELECT started,ended,reason FROM incidents WHERE site=? AND started>? ORDER BY started DESC LIMIT 30')225 .all(app, Date.now() - 30 * 24 * 3600_000);226 const errors = db.prepare('SELECT ts,code,err,ms FROM checks WHERE site=? AND ok=0 ORDER BY ts DESC LIMIT 20').all(app);227 // uptime quotidien 14 jours pour la barre228 const days = [];229 for (let i = 13; i >= 0; i--) {230 const d0 = new Date(); d0.setHours(0, 0, 0, 0);231 const start = d0.getTime() - i * 24 * 3600_000;232 const r = db.prepare('SELECT COUNT(*) n, SUM(ok) up FROM checks WHERE site=? AND ts>=? AND ts<?').get(app, start, start + 24 * 3600_000);233 days.push({ day: start, uptime: r.n ? Math.round((r.up / r.n) * 1000) / 10 : null });234 }235 return { app, label: s.label, domain: s.domain, node: s.node, dir: s.dir, series, incidents, errors, days };236}237238export function ecoIncidents() {239 return db.prepare(`SELECT site,started,ended,reason FROM incidents ORDER BY started DESC LIMIT 50`).all();240}241242export function ecoNodes() {243 const out = [];244 for (const node of NODES) {245 const last = db.prepare('SELECT ts,load1,mem_used,mem_total,disk_pct FROM nodestats WHERE node=? ORDER BY ts DESC LIMIT 1').get(node);246 const hist = db.prepare('SELECT ts,load1,mem_used,mem_total FROM nodestats WHERE node=? AND ts>? ORDER BY ts').all(node, Date.now() - 24 * 3600_000);247 const apps = SITES.filter(s => s.node === node).map(s => {248 const p = db.prepare('SELECT cpu,mem,status,restarts FROM appstats WHERE app=? ORDER BY ts DESC LIMIT 1').get(s.app);249 return { app: s.app, label: s.label, ...p };250 });251 out.push({ node, last: last || null, hist, apps });252 }253 return { nodes: out, ts: Date.now() };254}255