The benchmarking process monitors two key CPU usage metrics: App CPU Usage and Database CPU Usage. These metrics represent the average percentage of CPU resources utilized by the application (App CPU Usage) and database server (Database CPU Usage) during peak hours.

They are calculated as follows:

• App CPU Usage Percentage: (Average number of used OCPUs across all OKE nodes during peak hours) / (Total number of OCPUs per node) x 100%
• App CPU Usage (Cores Consumed): (Application CPU Usage percentage) x (Total Number of Worker Nodes) x (Number of Cores)
• Database CPU Usage Percentage: (Average number of used CPU cores across both database nodes during peak hours) / (Total number of CPU cores per database instance) x 100%
• Database CPU Usage (Cores Consumed): (Database CPU Usage percentage) x (Total Number of RAC Nodes) x (Number of Cores per Node)

Throughput (ops/sec) per Schema is a metric that reflects the transaction processing speed for each individual schema within the system. It essentially measures the number of transactions processed per second for a specific schema. It is calculated as follows:

Throughput (ops/sec) per Schema = (# of Accounts in a Schema) / (Processing Time per Account for that Schema)

They are calculated as follows:

• Overall Throughput Efficient Index: This is calculated by dividing the Total Throughput (op/sec) by the sum of App oCPU consumed and DB oCPU consumed. This index provides an overall system efficiency rating, encompassing both application and database resource utilization.
• App Throughput Efficient Index: This metric focuses specifically on application efficiency. It's calculated by dividing the Total Throughput (ops/sec) by the App oCPU consumed alone. Analyzing this index helps identify how efficiently the application utilizes CPU resources for processing transactions.
• Database Throughput Efficient Index: Similar to the App Throughput Efficient Index, this metric focuses on database efficiency. It's calculated by dividing the Total Throughput (ops/sec) by the DB oCPU consumed to assess how effectively the database server utilizes its CPU resources for processing transactions.

The benchmarking process also evaluates input/output operations per second (IOPS) efficiency through three metrics: IOPS Overall Efficiency, IOPS Read Efficiency, and IOPS Write Efficiency.

They are calculated as follows:

• IOPS Overall Efficiency: This metric is calculated by dividing the Total Throughput (ops/sec) by the combined Read IOPS and Write IOPS. It provides an overall picture of IOPS efficiency for all system operations.
• IOPS Read Efficiency: This metric focuses on read operations specifically. It is calculated by dividing the Total Throughput (ops/sec) by the Read IOPS, offering insights into how efficiently the system processes read requests.
• IOPS Write Efficiency: Similar to IOPS Read Efficiency, this metric focuses on write operations. It's calculated by dividing the Total Throughput (ops/sec) by the Write IOPS, providing insights into how efficiently the system handles write requests.

Latency refers to the time taken by the system to process a transaction. The 95th percentile latency is a metric that indicates the time taken to process 95% of transactions within the system.

Memory usage is monitored for both the application (App Memory Usage) and the database server (DB Memory Usage) during peak hours.

They are calculated as follows:

• App Memory Usage: (Average used memory per OKE node during peak hours) x (Total number of OKE nodes)
• Database Memory Usage: (Average used memory per database instance during peak hours) x (Number of database instances)

Some common metrics captured during testing for the Billing Care REST API include:

• Response Times: Time taken by the API to respond to each request. Analyzed to identify potential bottlenecks.
• Throughput (ops/sec): Number of API requests processed per second. Provides an overall picture of processing speed under load.