Uptime checks
External checks on the routes that matter — not only the homepage — run from more than one region.
Monitoring & alerts
Know it broke before your customers tell you
The worst way to learn about an outage is a support email. Monitoring is a couple of days of work and it changes how the whole thing feels to run.
In short
Monitoring means three things running at once: uptime checks that confirm the app is reachable, error tracking that captures every exception with its stack trace and the session that caused it, and performance alerts that fire when latency or error rate crosses a threshold you set. Gen2Prod wires all three to Slack or email with thresholds tuned from a week of your real traffic — because an alert people learn to ignore is worse than no alert at all.
- Typical timeline
- 2–3 days
- Alerts to
- Slack, email or PagerDuty
- Covers
- Uptime, errors, latency and logs
- Available as
- A one-off setup or an ongoing retainer
You probably need this if…
- You found out about the last outage from a customer.
- Errors go to the browser console and nowhere else.
- There are no logs from yesterday because nothing kept any.
- You can't answer “was it slow last Tuesday?”
- Alerts exist, and everyone has muted the channel.
What monitoring & alerts covers
Error tracking
Every exception captured with a stack trace, a release version, and the session that caused it, mapped back to your source.
Latency & performance alerts
Thresholds on response time, error rate and Core Web Vitals, tuned so a page firing means something is genuinely wrong.
Logs you can search
Structured logging with retention, so “what happened at 3am” is a query rather than a guess.
A dashboard worth opening
One screen — traffic, errors, latency and the last deploy — enough to answer “is it fine?” in five seconds.
Alert routing
Who gets woken, for what, and how it escalates: agreed with you and written down.
How we run it
- 1
Instrument
Error tracking, tracing and structured logs added to the app and tied to your release versions.
- 2
Tune the thresholds
Set from a week of your real traffic, so alerts fire on incidents rather than on Tuesdays.
- 3
Rehearse an alert
We trigger one end to end, so you know the notification arrives before the night you need it to.
What you get back
- Uptime, error and performance monitoring live
- Alerts routed to Slack or email, thresholds tuned and tested
- A dashboard covering traffic, errors, latency and deploys
- An incident runbook: what to check first, how to roll back
Tools we reach for
- Sentry
- Uptime monitoring
- Structured log storage
- Grafana
- OpenTelemetry
- Slack & PagerDuty
- Platform analytics
Monitoring & alerts — questions we get asked
What does monitoring cost to run?
Most apps at early traffic sit inside the free or entry tiers of an error tracker and an uptime service — nothing to a few tens of dollars a month. We size the tools to your traffic and tell you where the next price step is.
Can you watch it for us?
Yes — that's the monitoring retainer. We take the alerts, triage them, and fix what falls to us. The pricing page sets out what's included.
Does this add user tracking to my app?
No. Error and performance monitoring watches your servers and your code, not your users' behaviour. We configure it to scrub personal data by default and tell you exactly what leaves your app.
We already have Sentry. Is that enough?
It's a good start, and it's usually installed but not tuned: no release tracking, no source maps, no alert rules, and thousands of noise events nobody reads. Making an existing setup genuinely useful is often a single day of work.
Often paired with
All servicesAI feature build-out
The demo worked. Then it needed streaming, retries, cost caps, and an answer for what happens when the model is wrong. That's the part we build.
Learn moreHandoff & docs
The goal isn't to keep you on a retainer. It's to hand back something a new engineer can be productive in on day one.
Learn moreRefactor & cleanup
Lovable, Bolt, v0 and Cursor write code fast — and they write it flat. We reshape it into modules with clear boundaries, so the next feature takes an afternoon instead of a week.
Learn more