Problems using Redis with Wordpress on GCP

Hello forum, I ask for help here on the forum about integrating Redis with WordPress.

I’m trying to use Redis in an instance on GCP with MemoryStore to optimize WordPress by integrating with the Redis Object Cache plugin in the application, and during load tests, I’m not able to move forward because the connection with Redis is interrupted with few accesses. tell me some reference that can configure and study?

The WordPress application workload has separate resources such as vm for the application (4vCPU and 8GB), CloudSQL instance for the mysql database, and a CloudStorage bucket to store the uploads, all in order to eliminate contention for resources within from vm.

For this POC, I’m testing the following Redis configurations:

Instance settings in MemoryStore:
Capacity: 1GB
version: 6.x
Replicas read: 2

Behavior information:

I perform a load test with the ApacheBenchmark Tool, and when the application receives 9 simultaneous requests, the connection of the instance with Redis is interrupted.

Extra information: when the Redis service is embedded in the vm, the same behavior occurs, but the Redis service consumes a lot of memory until it disconnects from the application, a difference is that it supports many more requests than when it is in an instance in the GCP.

image

Greetings from Brazil to all, sorry if the question is outside the scope of the chosen category.