Based on the available sources, I don't have specific documentation about the

pool_max_idle_per_host

parameter or configuration options to modify it for achieving higher concurrent user loads. (1) (2) (3) The "ERROR cannot send to opa" message you're encountering when scaling to 100 CCU suggests a connection pooling or timeout issue between the PDP and OPA, but the available documentation doesn't provide guidance on adjusting this specific parameter. For achieving 100-200 CCU per PDP, the documentation does mention several relevant configurations that might help: ​*PDP_OPA_CLIENT_QUERY_TIMEOUT*​ - Controls the timeout for querying OPA for an allow decision (default is 1 second for version 0.9.0+) . (1) ​*PDP_OPA_CLIENT_FAILURE_THRESHOLD*​ - The percentage of failed requests to OPA that will trigger a failure (default is 0.1) . ​*OPAL_FETCHING_WORKER_COUNT*​ - The number of workers to use for fetching data concurrently (default is 6) . (1) (3) However, I don't have information about

pool_max_idle_per_host

or how to configure HTTP connection pooling parameters for the PDP-to-OPA communication. To help investigate this issue, could you provide: 1. Which Workspace / Project / Env you are using 2. What calls you are performing (as SDK code or CURL) 3. Which PDP version you are running 4. Relevant logs from your PDP showing the "ERROR cannot send to opa" messages This will help the engineering team investigate the connection pooling behavior and provide guidance on the appropriate configuration.

To learn more, see the sources I considered: [[1] Configurations](https://docs.permit.io/concepts/pdp/configuration) [[2] Cloud PDP Capabilities](https://docs.permit.io/concepts/pdp/cloud-pdp-capabilities) [[3] OPAL Configuration Guide | OPAL](https://docs.opal.ac/getting-started/configuration)