Introducing Monitoring — Servers, Uptime and Your Backend in One Place

Your users find out first

An app that looks broken is usually a backend that is broken. The disk filled up, a container restarted in a loop, a deploy made one endpoint slow, or an exception started firing on every request. The first you hear of it is a one-star review or a feedback message saying "nothing loads".

SupportDock already holds the feedback. Now it watches the backend too, so you can catch the problem before the message arrives.

What it watches

Servers and containers. A small agent (one static Go binary, no dependencies) reports CPU, memory, disk, network and every Docker container on the host every 15 seconds. If the network drops, it buffers an hour of data and sends it once it can reconnect.

Uptime. HTTP(S) and TCP checks run from outside your network. You get response times, uptime percentages, and a warning before your TLS certificate expires.

Your backend (NestJS). The SupportDock SDK times every request by route, picks out slow database queries, and groups errors by fingerprint. Each app gets three new tabs:

• Health: requests, error rate and p95 latency at a glance
• Performance: your slowest routes and the queries behind them
• Errors: grouped exceptions with stack traces, first and last seen, and the release they appeared in

PHP apps. The PHP SDK reports caught exceptions, uncaught exceptions and fatal errors such as running out of memory, grouped the same way on the Errors tab.

Alerts that don't cry wolf

Sensible rules are switched on as soon as data arrives, so you don't have to set anything up:

• Server CPU, memory or disk too high, or a server that stops reporting
• A container that keeps restarting
• An uptime check failing twice in a row, or a certificate expiring within 14 days
• Error rate above 5%, p95 latency more than twice its usual level, or an error you have never seen before

A rule fires only when the condition holds for the whole window, and it resolves only once the whole window is clear, so a single spike doesn't page you. Alerts go to Telegram, email, or a webhook in Slack or Discord format. Quiet hours hold back everything except critical alerts.

Set it up

A server: in the console, go to Monitor → Add server and run the command it gives you:

curl -fsSL https://supportdock.io/monitor/install.sh | sudo SUPPORTDOCK_SERVER_TOKEN=sd_srv_your_token sh

A NestJS backend: install the SDK with its OpenTelemetry packages (the app's Health tab lists the full command), then start monitoring before anything else loads:

// src/instrument.ts: import it first in main.ts
import { initNestMonitoring } from '@bangkeut-technology/supportdock-sdk/nestjs';

initNestMonitoring({
    projectKey: process.env.SUPPORTDOCK_KEY!, // the same API key as feedback
    release: process.env.GIT_SHA,
    monitoring: { traces: true, errors: true, crashes: true, sampleRate: 0.2 },
});

It uses the same API key your app already uses for feedback. Query strings and SQL values are removed before anything leaves your server, and SupportDock removes them again on arrival.

A PHP app: composer require bangkeut-technology/supportdock-sdk:^0.6, then register the handlers as early as possible:

use SupportDock\Monitoring\Monitor;

$monitor = new Monitor([
    'projectKey' => getenv('SUPPORTDOCK_KEY'),
    'release' => getenv('GIT_SHA') ?: null,
]);
$monitor->registerHandlers();

Request timing and slow queries need NestJS for now. PHP apps get error tracking. Expo and Next.js are next.

Included on every plan

Free covers 2 servers and 5 uptime checks, with 30 days of history. Pro raises that to 20 servers, 50 checks and a year of history. Open Monitor to add your first server.