57dc91585d
Internal SmartGift build of a Claude Code monitoring dashboard. Lanes: a durable unit of parallel agent work, one per working directory, tracked across session restarts. Managed lanes are git worktrees the dashboard provisions and can reset or remove behind a three-check destroy guard and a counted preflight; adopted lanes are directories you already own and are never destroyable. Pipelines: a lane moves through pipeline stages. A stage the agent declares with evidence renders green; a stage inferred from the tool-event stream renders dashed amber and never counts as done. Detection is forward-only within a 30-minute window, and never writes the declared stage. Workspace: one page at /run with a lane grid, the selected lane's pipeline, and a full Claude console behind a disclosure.
201 lines
6.9 KiB
JavaScript
201 lines
6.9 KiB
JavaScript
/**
|
|
* @file metrics.js
|
|
* @description Prometheus / OpenMetrics text-exposition endpoint (GET /api/metrics)
|
|
* so this monitoring dashboard can itself be scraped into Prometheus / Grafana.
|
|
*
|
|
* The dashboard already tracks everything an operator wants on a wall board —
|
|
* live sessions, agent states, event throughput, token burn, connected realtime
|
|
* clients — but only over its own websocket + REST surface. This route re-exposes
|
|
* those same counters in the standard Prometheus text format (v0.0.4) so they can
|
|
* flow into an existing observability stack alongside the rest of a team's infra.
|
|
*
|
|
* All values are read straight from the same prepared statements the REST API
|
|
* uses (`server/db.js`), so the numbers line up exactly with the UI. The endpoint
|
|
* is READ-ONLY, allocates nothing persistent, and — being mounted under `/api` —
|
|
* sits behind the same guards as every other route: the DNS-rebinding Host-header
|
|
* guard and the optional `DASHBOARD_TOKEN` guard. A scraper that reaches the
|
|
* server as anything other than loopback (e.g. Prometheus in Docker via
|
|
* `host.docker.internal`) must therefore be allowlisted with
|
|
* `DASHBOARD_ALLOWED_HOSTS` (and send the token when one is set), so an instance
|
|
* never leaks operational data to an unexpected origin. The turnkey Prometheus +
|
|
* Grafana bundle in `monitoring/` documents the exact setup.
|
|
*
|
|
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
|
|
*/
|
|
|
|
const { Router } = require("express");
|
|
const { stmts, db } = require("../db");
|
|
const { getConnectionCount } = require("../websocket");
|
|
|
|
const router = Router();
|
|
|
|
// Resolved once at load — the dashboard version, surfaced as a build-info label
|
|
// (the canonical Prometheus idiom for exposing a version string as a metric).
|
|
const APP_VERSION = (() => {
|
|
try {
|
|
return require("../../package.json").version || "0.0.0";
|
|
} catch {
|
|
return "0.0.0";
|
|
}
|
|
})();
|
|
|
|
// Statuses are enumerated (not just whatever the DB currently holds) so a metric
|
|
// series never silently disappears when its count hits zero — a gauge that drops
|
|
// out of the exposition breaks rate()/alerting downstream.
|
|
const SESSION_STATUSES = ["active", "completed", "error", "abandoned"];
|
|
const AGENT_STATUSES = ["working", "waiting", "completed", "error"];
|
|
|
|
/** Escape a Prometheus label value (backslash, double-quote, newline). */
|
|
function escapeLabelValue(value) {
|
|
return String(value).replace(/\\/g, "\\\\").replace(/\n/g, "\\n").replace(/"/g, '\\"');
|
|
}
|
|
|
|
/**
|
|
* Render one metric family (HELP + TYPE header, then one line per sample) into
|
|
* the `out` line array. Each sample is `{ value, labels? }`. Non-finite values
|
|
* are coerced to 0 so a bad read can never emit an unparseable exposition line.
|
|
*/
|
|
function appendMetric(out, name, help, type, samples) {
|
|
out.push(`# HELP ${name} ${help}`);
|
|
out.push(`# TYPE ${name} ${type}`);
|
|
for (const sample of samples) {
|
|
const labels = sample.labels
|
|
? "{" +
|
|
Object.entries(sample.labels)
|
|
.map(([k, v]) => `${k}="${escapeLabelValue(v)}"`)
|
|
.join(",") +
|
|
"}"
|
|
: "";
|
|
const value = Number.isFinite(sample.value) ? sample.value : 0;
|
|
out.push(`${name}${labels} ${value}`);
|
|
}
|
|
}
|
|
|
|
// GET /api/metrics — Prometheus exposition of the dashboard's live counters.
|
|
router.get("/", (_req, res) => {
|
|
const out = [];
|
|
|
|
appendMetric(out, "ccam_up", "1 when the dashboard API is serving this scrape.", "gauge", [
|
|
{ value: 1 },
|
|
]);
|
|
appendMetric(
|
|
out,
|
|
"ccam_build_info",
|
|
"Dashboard build info; the value is always 1, the version rides on the label.",
|
|
"gauge",
|
|
[{ labels: { version: APP_VERSION }, value: 1 }]
|
|
);
|
|
appendMetric(
|
|
out,
|
|
"ccam_process_uptime_seconds",
|
|
"Uptime of the dashboard server process in seconds.",
|
|
"gauge",
|
|
[{ value: Math.round(process.uptime()) }]
|
|
);
|
|
appendMetric(
|
|
out,
|
|
"ccam_process_resident_memory_bytes",
|
|
"Resident set size (RSS) of the dashboard server process in bytes.",
|
|
"gauge",
|
|
[{ value: process.memoryUsage().rss }]
|
|
);
|
|
|
|
// Sessions by status.
|
|
const sessionCounts = new Map(stmts.sessionStatusCounts.all().map((r) => [r.status, r.count]));
|
|
appendMetric(
|
|
out,
|
|
"ccam_sessions",
|
|
"Number of sessions by lifecycle status.",
|
|
"gauge",
|
|
SESSION_STATUSES.map((status) => ({
|
|
labels: { status },
|
|
value: sessionCounts.get(status) || 0,
|
|
}))
|
|
);
|
|
|
|
// Agents by status.
|
|
const agentCounts = new Map(stmts.agentStatusCounts.all().map((r) => [r.status, r.count]));
|
|
appendMetric(
|
|
out,
|
|
"ccam_agents",
|
|
"Number of agents (main + subagents) by status.",
|
|
"gauge",
|
|
AGENT_STATUSES.map((status) => ({ labels: { status }, value: agentCounts.get(status) || 0 }))
|
|
);
|
|
|
|
// Event throughput (monotonic — a counter).
|
|
appendMetric(
|
|
out,
|
|
"ccam_events_total",
|
|
"Total hook and synthetic events recorded since the database was created.",
|
|
"counter",
|
|
[{ value: stmts.countEvents.get().count }]
|
|
);
|
|
|
|
// Connected realtime (WebSocket) clients.
|
|
let clients = 0;
|
|
try {
|
|
clients = getConnectionCount();
|
|
} catch {
|
|
/* websocket not up (e.g. under test without a server) — report 0 */
|
|
}
|
|
appendMetric(
|
|
out,
|
|
"ccam_websocket_clients",
|
|
"Currently connected realtime (WebSocket) dashboard clients.",
|
|
"gauge",
|
|
[{ value: clients }]
|
|
);
|
|
|
|
// Remote Data Sources, split by whether background auto-sync is enabled.
|
|
let enabledSources = 0;
|
|
let totalSources = 0;
|
|
try {
|
|
const rows = stmts.listRemoteSources.all();
|
|
totalSources = rows.length;
|
|
enabledSources = rows.filter((r) => r.enabled).length;
|
|
} catch {
|
|
/* remote_sources table absent on a very old DB — report 0 */
|
|
}
|
|
appendMetric(
|
|
out,
|
|
"ccam_remote_sources",
|
|
"Configured Remote Data Sources, split by auto-sync enabled state.",
|
|
"gauge",
|
|
[
|
|
{ labels: { enabled: "true" }, value: enabledSources },
|
|
{ labels: { enabled: "false" }, value: totalSources - enabledSources },
|
|
]
|
|
);
|
|
|
|
// Cumulative token usage by kind (baseline_* preserves pre-compaction totals,
|
|
// matching how the pricing endpoints total usage).
|
|
const tokens = db
|
|
.prepare(
|
|
`SELECT
|
|
COALESCE(SUM(input_tokens + baseline_input), 0) AS input,
|
|
COALESCE(SUM(output_tokens + baseline_output), 0) AS output,
|
|
COALESCE(SUM(cache_read_tokens + baseline_cache_read), 0) AS cache_read,
|
|
COALESCE(SUM(cache_write_tokens + baseline_cache_write), 0) AS cache_write
|
|
FROM token_usage`
|
|
)
|
|
.get();
|
|
appendMetric(
|
|
out,
|
|
"ccam_tokens_total",
|
|
"Cumulative token usage across all sessions, by kind.",
|
|
"counter",
|
|
[
|
|
{ labels: { kind: "input" }, value: tokens.input },
|
|
{ labels: { kind: "output" }, value: tokens.output },
|
|
{ labels: { kind: "cache_read" }, value: tokens.cache_read },
|
|
{ labels: { kind: "cache_write" }, value: tokens.cache_write },
|
|
]
|
|
);
|
|
|
|
res.set("Content-Type", "text/plain; version=0.0.4; charset=utf-8");
|
|
res.send(out.join("\n") + "\n");
|
|
});
|
|
|
|
module.exports = router;
|