How To Set the Appropriate Requests for the RealTheory Collector
- 27 Mar 2025
- 1 Minute to read
- Contributors
- Print
- DarkLight
How To Set the Appropriate Requests for the RealTheory Collector
- Updated on 27 Mar 2025
- 1 Minute to read
- Contributors
- Print
- DarkLight
Article summary
Did you find this summary helpful?
Thank you for your feedback
Background
The RealTheory Collector is designed to operate efficiently while using less than 0.25% of the cluster's CPU and memory resources. As the performance of the RealTheory Collector is impacted by the size of the cluster where the Collector is installed, the CPU and memory requests configured for the realtheorycollector pod are typically determined by the approximate vCPU count of the cluster. Larger clusters must have higher requests to ensure that the Collector has the resources it needs to function efficiently.
For some clusters, such as clusters that use cluster-level autoscalers, clusters with a high density of namespaces (greater than 100), and clusters with a high density of workloads (greater than 5000), the requests should be set higher to ensure acceptable performance. You can set higher requests by specifying a higher vCPU count when you generate the deployment manifest.
Solution
When you create the deployment manifest for the RealTheory Collector, use the following guidelines when you specify the vCPU count:
| Cluster Characteristics | Calculate vCPU count using | Example |
|---|---|---|
| Typical | Approximate vCPU count | If the approximate cluster vCPU count is 500, set the vCPU count to 500 |
| High namespace density | Approximate vCPU count + 25-50% | If the approximate cluster vCPU count is 500, set the vCPU count somewhere between 625 and 750* |
| High workload density | Approximate vCPU count + 25-50% | If the approximate cluster vCPU count is 500, set the vCPU count somewhere between 625 and 750* |
| Uses autoscalers | Highest 30-day vCPU count | If the highest 30-day vCPU count was 780, set the vCPU count to 780 |
*The higher vCPU count setting does not represent an increase in the vCPUs on the target cluster. It is the configuration setting to specify when you generate the deployment manifest in the RealTheory console or the RealTheory API to ensure that the appropriate CPU and memory requests are set.
For more information on how to generate the RealTheory Collector deployment manifest, including how to configure the vCPU count setting, see the following:
Using the RealTheory Console:
Installing RealTheory in a Cluster
Using the RealTheory API:
How To: Generate the Collector Deployment Manifest
Was this article helpful?
