Uptime monitoring for SaaS and startups
Watch five things: the marketing site, the app, an API health endpoint, the login page and your scheduled jobs. Check the app and API every 30 seconds, treat a slow health endpoint as down, and put a status page in front of customers so an outage is one announcement instead of forty support tickets.
What breaks in a SaaS product, and what sees it
A SaaS outage is rarely "the server is off". It's the API returning 502 while the marketing site is fine, the login page timing out because the session store is full, or the nightly billing job silently not running for a week. Each needs its own check, because each fails on its own.
Every failure is confirmed by a second check before anyone is alerted, so a single slow response doesn't wake you up. Alerts say what was seen, for example "api.example.com/health answered 503" or "login timed out after 10 seconds", so the person on the other end of the Slack message knows where to look first.
What to monitor
| Page / URL | Check | Why |
|---|---|---|
| Marketing site (example.com) | Status 200, every 60 s | Signups come from here; also where customers look first when something is wrong. |
| App (app.example.com) | Status 200 or expected redirect to login, every 30 s | The front door for paying users. |
| API health (api.example.com/health) | Status 200 and body contains "ok", every 30 s | A health endpoint that checks the database and cache tells you about back-end trouble before users do. |
| Login page | Status 200 and contains "Sign in" | Auth providers and session stores fail separately from the app. |
| Webhook receiver (/webhooks/stripe) | Expect 405 or 400 on GET (your own rule), every 60 s | Proves the route is reachable; a 502 here means missed payment events. |
| Billing / cleanup cron jobs | Heartbeat: ping after each run, grace 15 min | The job that didn't run is invisible until the invoices don't go out. |
| Certificate and domain | Automatic reminders at 14/7/3/1 and 30/14/7/1 days | api.example.com has its own certificate; forgetting it takes every integration down. |
Cron jobs and queue workers
Scheduled jobs are the quiet failure. A heartbeat monitor gives you a URL; the job calls it when it finishes successfully (add && curl -fsS https://heyuptime.com/ping/<token> to the crontab line). You choose the period (every 5 minutes to every week) and a grace period. If the ping doesn't arrive in time, you're alerted that the job is late, with the time it was last seen.
Use it for billing runs, digest emails, report generation, data syncs and the queue worker's own "I'm alive" ping. Each heartbeat counts as one monitor.
A status page customers can subscribe to
A public status page with App, API and Website as components turns an outage from a flood of "is it just me?" tickets into one place to point people. Customers subscribe by email and get the outage notice and the all-clear. The page shows 90 days of daily history, so a prospect doing due diligence can see the record rather than take your word for it.
Starter includes one status page, Pro five, Business twenty. Maintenance windows show on the page as planned work, not outages.
Alerts where the team already is
- Email, Slack, Discord, Microsoft Teams, Telegram and webhooks, all on every plan, so alerts can open an incident in your own tooling.
- Pick channels per monitor: API and app to the engineering channel, the marketing site to the founder's email.
- The weekly summary gives uptime and response time per monitor, which is the number investors and enterprise prospects ask for.
Which plan for a startup
In short: Pro ($12/month: 50 monitors, 30-second checks, 5 status pages, a year of history) fits most startups; Business ($29) when you need more than 50 monitors or more status pages.
The 14-day trial needs no card and includes every feature. Monitoring is outside-in HTTP(S) checks, heartbeats, certificate, domain and DNS watching; there is no agent to install and nothing to deploy.
Questions
Should I monitor the API or the app?
Both, as separate monitors. The API's health endpoint catches back-end problems (database, cache, queue) that the app's front page hides behind a loading spinner. The app monitor catches the front end, CDN and auth being broken while the API is fine.
What should a health endpoint return?
A 200 with a short body such as {"status":"ok"} only when the database and anything else essential responded. Add a rule that the body must contain "ok" so a 200 with an error message still counts as down.
How do I monitor a cron job or background worker?
Create a heartbeat monitor, which gives you a URL. Have the job call it after each successful run. Set the period to the job's schedule and a grace period; if the call doesn't arrive in time, you're alerted that the job is late.
Does heyUptime check from several locations or run browser tests?
No. Checks are plain HTTP(S) requests from heyUptime's servers, confirmed by a second check before alerting, plus heartbeats, certificate, domain and DNS watching. There are no browser-based checks, server agents or multi-region probes.
See it on your own site.
heyUptime checks every 30 seconds, confirms before alerting, and tells you why it's down in plain words. Free for 14 days, no card.
Related
- Uptime monitoring for indie developers
- Cron job monitoring
- Status page example
- What 99.95% uptime means
- Pricing
- Uptime monitoring for online shops
- Uptime monitoring for web agencies
- Uptime monitoring for WordPress sites
- Uptime monitoring for Shopify stores
- Uptime monitoring for freelancers and small studios
- Uptime monitoring for side projects and indie developers
Updated October 8, 2026.