We spend a lot of time building automation, tooling, and monitoring to equip developers to understand and optimize their datastore usage and minimize the impact of any operational issues on customers and developers alike. For example, we’ve built database clients to make monitoring and analysis of usage patterns easier for HubSpot teams and are working to commit these changes back to the community to better support and monitor in a multi-tenant environment.
We’re also operationally responsible for a huge volume of traffic to and from these datastores. Our HBase clusters serve over 3 million requests/second across 220+ tables, while our ElasticSearch clusters serve over 20k searches/second and 50k indexes/second to 90+ billion documents. Streaming that data to and from applications amounts to more than 3 GB/sec of data through our Kafka clusters, with hundreds of producers and consumers
What we’re looking for