Memorystore Redis slow resize

Google Cloud Database
1

Hi,

We are running a 1GB redis memorystore 6.x instance as a cache for a website. Today it was under a lot of pressure, and I resized it to 2GB. The resize took more than 1 hour, which was unexpected. GCP cannot be certain what situation a customer is facing. It might be downtime, and so should be taken seriously. Here are ideas:

  1. At the time of resize, make an estimate about the time it will take. Show that information to the user. Suggest it might be quicker to just launch a new memorystore instance.

  2. Offer more options, for example “If you are willing to shut-off or reboot the instance, which does create downtime, we can resize in 5 minutes. Otherwise the live on-line update will take 1 hour.” Give the user these choices. Maybe the 5 minutes of downtime is worth it. I would have chosen that.

  3. Of course try to see why a resize is slow from a technical perspective and make this go faster.

Thanks.

2

Thanks for your feedback. let me follow up on this for you.