This page describes product constraints that you should take into consideration
when planning and implementing your Memorystore for Redis solution.

## Replication backlog for Standard Tier instances

Standard Tier Memorystore for Redis instances reserve 10% of primary-node
memory for replication backlog usage. When [monitoring your instance](https://docs.cloud.google.com/memorystore/docs/redis/monitoring-instances),
this backlog appears in the **used memory** metric, even if the instance is
empty.

## You can only run one admin operation at a time

Examples of Memorystore for Redis admin operations are resizing, renaming, and
changing other Redis instance configurations. Only one of these commands can be
run at a time. You will have to wait for any admin operation to complete before
you can run another one. During these operations, you can still run Redis
commands and access the data in the Redis instance.

## Quotas and limits

For information on the Memorystore for Redis instance capacity quota and API
requests quota, see [Quotas and Limits](https://docs.cloud.google.com/memorystore/docs/redis/quotas.md.txt).