Redis Server does not seem to scale correctly

Question

In our stage (pre production) environment, we suddenly (December 18th at 8:30 pm ET) started having errors with the Redis cache. Our production environment which is similar but scaled up is working fine and the deployed software is identical. The errors we are seeing are similar to this stack overflow, but not sporadically but all the time. If we reboot the redis instance they stop briefly and then restart. These errors do not show up in prod at all.

https://stackoverflow.com/questions/31328212/no-connection-is-available-to-service-this-operation-when-using-azure-redis-cac

The scale size in prod is a C5 Standard and the (overview) memory utilization is appropriate for that (about 23 GB). (the AI Used Memory reports 140 GB)

The scale size in stage was a C1 Standard but the (overview) memory utilization topped out at 100 MB.

I rescaled the stage instance to a C4 Standard, but, the (overview) memory utilization still is stuck at 100 MB. (the AI Used Memory reports 1 GB)

I tried rebooting the Redis instance, but that had no effect, still stuck at 100 MB.

Is there something else I need to do to get it to really use the new size? I've tried googling this, but, my google-fu is not helping today.