You may need to verify that the Logic App is set to scale automatically(check under the Workflow Settings in your Logic App. For the event hub, you can use a partitioned consumer model. More partitions mean more parallelism and higher throughput. Consider using Event Hubs Dedicated clusters if the expected load is very high.
Try also to increase the batch size to process more messages per trigger invocation. However, be mindful of the message size and the processing time.
You can use Azure Blob Storage with high throughput and low latency for large or intermediate data storage needs.
Each action in the Logic App can add latency so you may need to minimize the number of actions and use the parallel processing capabilities of Logic Apps where applicable. Don't forget that the long-running operations can block the processing of other messages.
You can use Azure Monitor and Application Insights for diagnostics and to identify performance bottlenecks.
If your Logic App interacts with other Azure services or external endpoints, consider network optimization strategies like Azure ExpressRoute for more predictable network performance.
For even higher throughput and performance, consider deploying your Logic App in an Integration Service Environment (ISE), which provides dedicated resources.
Here are some links to help :