The cost of scale and Grafana LGTM

Ubisoft - 2 Avenue Pasteur Saint Mandé, 94160 Paris
Thu, Sep 15, 2022, 7:00 PM (CEST)

Bonjour à toutes et à tous, C'est la rentrée pour le CNCF Paris aussi! Nous revenons avec un super événement et nous aurons le plaisir d'en apprendre plus sur la stack de Grafana ainsi que sur VictoriaMetrics et les problématiques lié au coût d'une base de données "at scale". À très bientôt, L'équipe Cloud Native Computing Paris

About this event

Note: L'événement aura lieu dans les locaux de Ubisoft et sera aussi diffusé en live et enregistré. Nous reviendrons vers vous prochainement avec plus de détails.

The cost at scale


Aliaksandr Valialkin, Co-Founder & CTO @ VictoriaMetrics

Aliaksandr is a seasoned CTO having worked in software development for almost two decades. Prior to building VictoriaMetrics, he was CTO of Pay Per Click affiliate network NextAdNet, where he worked to significantly optimize stats processing speed and reduce hosting cost. He holds a Master’s Degree in Computer Software Engineering from the Belarusian State University of Informatics and Radioelectronics and is a well-known author of popular performance-oriented libraries fasthttp, fastcache and quicktemplate.


When people choose a database, they tend to think about how it would scale, how far they can get with it. They read benchmarks, case studies, and look for big numbers. But what they could be missing is the price one needs to pay for scale.

The cost of scale defines not only hardware price, but also man-hours, maintenance burden, and complexity overhead. We'll discuss how we tried to solve this problem at VictoriaMetrics and how far we have come.

Cloud Native Observability with the LGTM stack: Loki, Grafana, Tempo & Mimir


Abdelkrim is a Solutions Engineering Manager at Grafana Labs with over a decade of experience on distributed and data intensive systems. Prior to Grafana Labs, he worked as a Solutions Engineer at Hortonworks, Cloudera and Imply. Abdelkrim is an open source advocate who loves sharing best practices with the community at meetups or tech conferences such as GrafanaCon, ApacheCon, or Flink Forward. He holds a PhD, MSc and MSe degrees in Computer Science.


Imagine if you can take the Prometheus design, amplify its strength, fix its limitations and generalize it to logs and traces. You get a full observability stack that is easy to use, scalable and integrates natively with K8S. That's what we did by building Mimir, Loki and Tempo.

In this talk, we will review the design principle of the LGTM stack, its architecture as well as its benefits (native correlation, alerting, unified architecture, low TCO, etc). We will explain why the LGTM stack is great for observing Cloud Native applications, from infrastructure to services. Finally, we will show you how to get started with it.


Thursday, Sep 15
7:00 PM - 9:45 PM (CEST)


2 Avenue Pasteur Saint Mandé94160