ClearDB Iron Performance Metrics Based on Live Workloads
ClearDB Iron is our high-speed database system that leverages MySQL (or any other relational database created in the last 30 years) to power extreme workload requirements that have tight performance variance SLA requirements. Recently, we began to closely audit the raw performance metrics coming out of one of our ClearDB Iron-based multi-tenant stamps in our EC2 based footprint to see what real-world workloads are pushing ClearDB Iron towards in terms of queries per second. The reason for wanting to monitor these metrics really stemmed from diagnostics and the need to properly account for the throughput that the system was producing. Under the covers, ClearDB places special software both above and beneath MySQL, as well as wraps the entire system with proprietary cluster software that we have built so that the cluster, and the system that manages it, is as close to fully autonomous as possible. This results in predictable operations as well as an overall better experience for our customers.
Since July of last year, we have noticed a substantial decrease in support issues related to performance from our EC2 based fleet, and the primary reason for this has clearly shown to be ClearDB Iron and how it handles I/O throughput. The sheer I/O processing and innovation investment that we have placed into providing a fantastic MySQL experience has been fairly significant, but based on the graph below, we believe that it has been well worth it. Please note that this graph is based on a sample set of metrics that was taken yesterday over the course of an hour.
We will continue to post interesting ClearDB Iron metrics like this as we continue to build upon and release ClearDB Iron’s availability in other operating environments in the future. Enjoy!