OCI Application Performance Monitoring runs over 500 instances of Autonomous Database with no dedicated DBAs, saving the equivalent of 30 admin jobs.
“With hundreds of databases in dozens of regions we have no other way than Oracle Autonomous Database to provision and monitor our databases. Patching, backups, and scaling just wouldn’t be possible with our team.”
Application users are very sensitive to performance. If a website or shopping cart application runs too slowly, customers will simply abandon their basket and go elsewhere. Internal users of corporate applications don’t have that option, but they can and will register their displeasure. In both cases, poor performance hurts core business processes and internal support groups need to correct the problem.
This is where OCI Application Performance Monitoring comes in. A component of Oracle Cloud Observability and Management Platform, it can monitor microservices-based applications as well as legacy multitier ones. It captures detailed telemetry from every step of every transaction in every application component. It can track overall performance, measure the experience for individual users, and also proactively run synthetic monitors to ensure availability. By identifying problems before users encounter them, users have been able to reduce application performance issues by up to 90%.
Why Oracle chose Autonomous Database
The data-intensive process includes capturing millions of records per hour during normal operations, while the largest environment handles over 1 billion records per day. And there are real time requirements as well: Users need to have analytics access to data within 30 seconds of it being generated. All that data needs to go somewhere, and the development team selected Oracle Autonomous Database as the destination.
Autonomous Database is transparent to users of the service. But it provides Application Performance Monitoring with essential capabilities to deliver better performance for customers as well as reduced cost and effort for Oracle Cloud operations teams.
Using Oracle Autonomous Database allowed the development team to move to a single-tenant architecture. By isolating tenants, each database instance can optimize itself for the needs of that particular user, increasing performance. It eliminates any potential “noisy neighbors” problem, simplifies the security certification process, and enables the development team to have much finer control of capacity for each individual customer.
There’s lots of upside to a single-tenant architecture, but the potential downside is there are many more databases to manage. In fact, based on the administration workload of the previous version, the application performance management team estimated that managing over 500 non-autonomous databases would likely take more than 30 full-time administrators. With Autonomous Database, however, that requirement falls to zero. Patching that could have required downtime in the past is now accomplished without bringing the system down at all.
However, even the best capacity planning can’t predict everything. Big workload changes have the potential to saturate any fixed database instance, so the application performance monitoring team switched on autoscaling. Any customer database can autonomously scale up by 3X to support transient changes and give the team time to increase the base number of OCPUs if the need becomes permanent. The team is also building a rule engine that detects sustained growth and will interact with the autonomous scaling to expand base resources automatically.