We Tested and Compared 6 Database Operators. The Results are In !
In the early days of Kubernetes, stateful workloads (like databases) were not often run on Kubernetes. Times have changed, and the stateful landscape has matured, with the CSI (Container Storage Interface) being supported by many cloud and storage vendors, and the DoK (Data on Kubernetes) community showing and enabling growing adoption of K8S for databases and stateful workloads. Running a database on Kubernetes is relatively straightforward, but running a production database is more complex: we want replication, automated failover, backups, PITR (point-in-time recovery), just to name a few. Many operators claim to take care of these tasks. CNPG, StackGres, MOCO, others from Zalando or Percona... But do they meet expectations? Our team has tested (and runs in production for our customers) most of these operators, and will give you a candid yet detailed review of our experience and how they stack up against each other in various scenarios.
@cncf
Speakers :
🏴☠️ Jérôme Petazzoni
🏴☠️ Alexandre Buisine
Suis-nous sur nos réseaux sociaux :
➡️ LINKEDIN : / enix-sas
🐥 TWITTER : / enixsas
😸 GITHUB : https://github.com/enix
🐘 MASTODON : @[email protected]
💫 WEB : https://enix.io/fr/