Choosing datastores

MySQL is used as a primary data store by a majority of internal tools at LinkedIn. We host over 2000 databases of different use-cases. To support our clients better, we have developed a few tools around MySQL. In this talk we will discuss how our tools improved the developer experience while using MySQL.
We will talk about how a database is requested and provisioned by the clients and how they get access to their respective databases the way they want i.e. from having an all-round admin access for complete control to only read-only access for Select queries. We will also throw some light on the Query Analyzer tool that has helped developers identify and fix their slow queries. Lastly, a brief about how the MySQL servers are highly available for the developers, which is achieved by the MySQL Orchestrator . With these tools in action, we provide the developers a managed MySQL experience at LinkedIn.

Key Points

• How these tools make life of developers easy when they deal with MySQL
• How access for MySQL are managed efficiently
• How QAN adds value
• How the applications feel negligible glitch when a primary host goes down and is failed over to a replica by orchestrator

Audience Takeaways

• How we created a Managed MySQL Experience at LinkedIn
• QAN is open sourced

Session Outline

• Introduction (1 min)
• MySQL at LinkedIn (3 min)
  • Basic Infrastructure
  • Different types of use-cases we support
• How a DB is provisioned on MySQL (3 min)
• Access Management (6 min)
• Query Analyzer (6 min)
• HA powered by MySQL Orchestrator (5 min)
• Workflow in Motion (3 min)
  • From provisioning to analysing queries
• Conclusion + Q&A (3 min)