Percona Live Online

It’s like picking a flavor of ice cream. You have your go-to databases, your favorites, the ones you know and love. But are they always the right choice for every task? And, the reality is you’re going to be working across five or six anyway. So how to choose? Instead of fitting the workload to the database, you fit the database to the workload. Is your workload mostly reads? Writes? Or a mixture? And what kinds of reads, linear scans, random reads? Do you need a database for transactioning or for durability? We’ll walk through the many open source options, project by project, and map workloads to the right database that can make or break the success of your next project.

Allegro is growing fast. Nobody sees it better than db admin. Environments with hundreds of users and their unique micro services, demanding over a dozen new databases every day, new cluster every week, and all that with maximum performance and resource isolation makes our job non trivial. I would like to present you a story of changing our architecture of bare bone physical servers setup for mongodb clusters, to fully automated, resource isolated architecture of micro clusters, based on statefulsets. How did we automate our processes, what obstacles we had to avoid, where we are today, and what are we gonna do next, to make mongodb + k8s even more flawless.