I have the following server specifications:
- 128GB RAM
- 2 CPUs with 12 cores each (24 cores) and a speed of 2.9 GHz
I currently have two services running on this server:
- The first service filters input data and directs the output to a streaming service, utilizing the 'choice element' for this task.
- The second service performs various tasks such as GeoTagging, incident indication, applying filters, and generating outputs.
Recently, I conducted a simulation using a CSV file containing 280,000 records, simulating real-time conditions with an average input rate of 25 inputs per second. Over the course of three days, the simulator sent approximately 2 million records to the input. Upon halting the simulation, I observed that the input retained the expected 2 million records. However, the two services only processed 1 million records (half of what was sent by the simulator). Despite this, the services are still operational, receiving input and producing output data.
The maximum rate of the services didn't exceed 5/sec
I have two questions regarding this scenario:
- Is it normal for the services to operate at half the speed of the input?
- How can I improve the performance of GeoEvent services to ensure real-time processing, considering that our system is crucial for emergency operations and requires truly live data without any delay?
Your guidance on addressing these concerns would be greatly appreciated. Thank you.
