Your fractional performance & reliability team.

We measure, we understand, we improve. Then we do it again.

A team of SRE and DevOps engineers providing managed services for your high-traffic web platforms. Operations, monitoring and 24/7 on-call, alongside your developers.
  • Traffic spike detected?
  • Response time critical?
  • Intermittent server errors?

If you are here, you are probably facing one of these problems

  • developers have to handle servers and incidents

    Incidents, maintenance, upgrades: that time comes out of product development.
  • performance and scaling are managed blind

    The site copes day to day, then degrades as traffic grows. Without reliable measurements, there is no telling where the bottleneck is.
  • changing the infrastructure safely is hard

    The infrastructure has grown over time. Changing it or deploying a new service causes incidents.
  • alerts come too late

    Too many or badly tuned, they fire once your users have already seen the problem. And finding the cause takes too long.
  • cloud costs keep creeping up

    Spending grows without anyone knowing what actually uses the resources.
  • backups have never been restored

    They exist, but nobody knows how, or how fast, the data could be restored after a major incident.

How we approach it

Performance is never settled for good. Traffic changes, code changes, infrastructure changes. So we work in a loop.
  1. Measure

    Monitoring, metrics, traces, logs. We look at what actually happens in production before making any assumption.
  2. Understand

    We trace the problem to its cause: a slow query, a setting, a capacity limit, or the code itself. When the problem is in the code, we read the code.
  3. Improve

    We fix it where it belongs, in the infrastructure or in the application with your developers. Then we check the effect on the numbers.
  4. Start again

    Each fix exposes the next weak spot. We measure again, and we prepare for the next traffic peak instead of reacting to it.

What we do

We work from the infrastructure up to the application code, on web platforms with very different stakes: e-commerce, SaaS, news and sports media, online gaming.
  • Monitoring and SLOs

    Metrics, dashboards, alerts that matter: every alert should call for an action. Together, we define SLOs (service level objectives): quality indicators, measured continuously, that your teams and ours track together.
  • Performance and scaling

    Response times, databases, caches. We get ready for traffic peaks before they hit.
  • Alongside your developers

    Help with debugging, optimisation, and code changes when that is where the problem lies. We also take part in design decisions.
  • 24/7 on-call

    Round-the-clock on-call for all our clients, handled by people who know your platform.
  • Costs under control

    Knowing what actually uses resources, sizing just right, and tracking spending over time.
  • Kubernetes

    Managed Kubernetes on AWS, GCP, Exoscale, OVHcloud and Scaleway. We also build and run clusters on your own servers with RKE2, along with their storage, with Rook and Ceph.

Experience from the field

Infrastructure designed to evolve, run day to day, and client relationships built for the long term.
  • 1 billion visits per month

    On a very high-traffic web platform we support.
  • 1 PB of distributed storage

    A Ceph cluster running in production.
  • 6 continents Infrastructure without borders

    Servers we run in about a dozen countries.
  • Over 10 years Clients who stay by choice

    Several of our clients have trusted us for more than ten years, with no minimum term. Their loyalty is earned day after day.

How we work

Fractional

An experienced team, shared between several clients. You get its skills for a fraction of the cost of a hire, and every client is covered by our 24/7 on-call. Some would call it a fractional SRE (Site Reliability Engineering) team.

With your teams

We are not a remote managed services provider carrying out decisions made elsewhere. We talk with your developers every day on a shared chat, and regularly over video calls. We work in your code when needed, and we have our say on the architecture.

No lock-in

Everything we set up is versioned in Git and shared with your teams, using open source tools: Debian, Kubernetes, Ceph, etc.
Nothing ties you to us: it’s yours.

You benefit from what we learn elsewhere

An optimisation found for one client is applied to the others. You get fixes for problems you have not run into yet.
  • Observability platform

    A shared platform, already proven on other critical platforms, with probes around the world. You don’t have to build it.
  • Kubernetes

    A common base, offered but never imposed. An improvement made for one of the clients using it is rolled out to all of them.

    Kers, our managed Kubernetes offering, is where it leads: an isolated virtual cluster per client, on shared infrastructure hosted in France.

  • Backups

    A proven solution, isolated from your environments and monitored by us. Restores are tested.

Get in touch

A project, or a question about your infrastructure: email or call us.

For our clients

Our clients reach us 24/7 by phone, on an on-call line reserved for them. Day to day, we talk with their teams on chat, and regularly over video calls: we are part of their team.

In an emergency, use the on-call line rather than email.