The 3am call comes to us.
Your developers keep shipping.
Critical Support gives software companies 24x7 support from our engineers. They go on call, nights and weekends included. When something breaks, they get it working again. Every month, they also fix the causes, so the next 3am call is less likely.
Where this fits: customers pay for your software, and whoever fixes it at night also ships in the morning.
The customer list grew. The on-call list didn't.
The 3am call goes to whoever built it.
The team is too small for an on-call rotation. So one developer takes the alerts at night, at weekends and on holiday. Nobody hired them for that.
A customer spots it before the alerts do.
Some alerts fire all night and mean nothing. Others stay quiet when something real breaks. So the first warning comes from a customer. Meanwhile, the monitoring bill goes up, and nobody can say why.
The fixes wait for a quiet week.
Feature work has a deadline. The alert fixes, runbooks and access reviews don't. So they wait for a quiet week, and the quiet week doesn't come.
Each of these is a sensible call on the day the team makes it. Together, they leave the nights resting on one person's goodwill. One more hire doesn't fix it, because one engineer can't cover every night, weekend and holiday.
On call is a team job, not a hero's.
Most teams start with the nights.
For most teams, the nights are the problem, so we start with Critical Support. Sometimes a different first step makes more sense. Wherever we start, it leads to a service our engineers run every month.
The route
We take the night calls first. Then we fix what causes them.
Our engineers go on call from the start. Each month, they also work through what keeps going wrong.
-
Usually first
Critical Support
Our engineers go on call 24x7. Under contract, they respond to SEV-1 and SEV-2 incidents within 15 minutes. Each month, they fix causes across six areas.
-
If you want the gaps found first
Cloud HealthScan
We check your AWS or Azure account, read-only, and rank what to fix first. It doesn't need Datadog.
Where the first step changes
If you run on one cloud and don't need every hour covered, start with Critical Support Lite. Our engineers cover the hours you choose, with fewer improvement hours each month. You can add 24x7 cover, or step up to Critical Support at any time.
If your own engineers already go on call and you want to keep that, start with Critical Response. It covers the hours they can't, in the window you choose. It handles incidents only. The monthly fixes come with Critical Support or Lite.
If the monitoring is the problem rather than the nights, we start with Datadog. That's usually too many alerts, problems that get missed, or a bill that keeps climbing. You don't need to have bought Datadog yet.
If you already run Datadog, Datadog HealthScan checks it, read-only. You get a scorecard and a ranked list of fixes. If you're still deciding on Datadog, FETCH puts our engineers on your trial at no cost.
If Datadog needs building or sorting out, LaunchPad rolls it out in three phases: assess, pilot and scale. Catalyst works through an agreed list of fixes. Our engineers do the work in both.
After that, Managed Observability looks after your Datadog every month: monitors, dashboards, tagging and the bill. This is the service we used to call Managed Datadog.
See every Datadog service, from the first trial to looking after it every month.
Many customers run Critical Support and Managed Observability side by side. Managed Observability looks after Datadog. Critical Support handles what breaks across your AWS and Azure accounts.
The nights stop landing on one person.
- With Critical Support, our engineers go on call 24x7. Under contract, they respond to SEV-1 and SEV-2 incidents within 15 minutes.
- Every incident goes through the same five stages: monitoring, triage, response, escalation, and recovery and review.
- Every SEV-1 incident gets a blameless postmortem. What it finds goes on the list of monthly fixes.
- Each month, our engineers fix causes across reliability, security, cost, performance, automation and governance.
- With Managed Observability, our engineers look after monitors, dashboards, tagging and the Datadog bill every month.
- You keep IAM and admin control. Material changes need your sign-off.
- You keep full access to your Datadog account at all times.
- Runbooks, standards and knowledge stay in your environment, not in one person's head.
Response time is a contractual commitment. Recovery time is a target, because recovery depends on the incident, not only on our speed.
We operate the stack. You own the product.
We fix it faster, then we fix the cause.
Critical Cloud reduces mean time to resolve incidents by 40 to 60 per cent.
24x7 cover and cleaner foundations after a move off Heroku
A UK operational excellence software company has its Azure platform watched 24x7, with every alert triaged and closed within SLA. Its engineers stay on the product.
Datadog onboarded in a trial window, saving a lean team many weeks
A UK media and entertainment software company used FETCH to try Datadog on a setup that looked like production, with expert help. It chose Datadog and skipped weeks of onboarding work.
Critical Cloud is the world's first Powered by Datadog accredited MSP, and a Datadog Advanced Partner. We're certified to ISO 27001 and Cyber Essentials Plus.
Common questions.
Do we keep control of our systems?
Yes. You keep IAM and admin control, access approvals, and every business and release decision. Material changes need your sign-off. You also keep full access to your Datadog account at all times.
Who approves changes?
You do, for every material change. Our engineers use Bits AI SRE and Watchdog to triage alerts and find likely causes. These tools only advise. A person approves every production, security and cost change.
Why not just hire an SRE?
One engineer covers about 1,856 of the 8,760 hours in a year, around 21 per cent. Nights, weekends and leave still need cover. Critical Support puts a team on call for around 75 per cent less than building one, at the standard of an in-house 24/7 team. The skills also stay when one person leaves. If you already run a staffed on-call rotation you want to keep, Critical Response covers the hours it can't.
Do we need Datadog?
Not to start. Cloud HealthScan doesn't need Datadog or any other monitoring tool. Critical Support, Critical Support Lite and Critical Response run on Datadog, in your own account. We set it up and look after it, and your team sees what we see. Datadog bills its licence on usage.
Where do we start?
Most teams start with Critical Support. If you want the gaps found first, we start with a Cloud HealthScan. If the monitoring is the problem rather than the nights, we start with Datadog HealthScan, or with FETCH if you're still on the trial. Tell us what's breaking on the first call, and we'll recommend one first step.
Someone has to answer at 3am. We'd rather it was us.
Tell us what you run and who answers the alerts at night. We'll recommend the simplest first step.