Updating our Redis instances from gp2 to gp3 disks

I’m looking at updating/respinning our Redis instances from gp2 to gp3 disks. I found the Aptible CLI command db:modify (aptible db:modify) which seems to do what I want and says it’s for “Databases which do not require downtime.” Is this the right thing to be looking at, and can you help give me a sense of what to expect here?

  1. Is Redis (with aof+rdb) a database which doesn’t require downtime?
  2. Is this safe to run on a database that’s currently in use, with active reads and writes?
  3. Is there any danger of data loss here?
  4. Is there any danger of additional load from this process competing with application needs?
  5. Any other gotchas to be concerned about?
  1. The modify command is supported for all databases, and does not require downtime for any database type. (Sorry for the confusion here, we made a grammatical error in our documentation which made it sound like only some database types would require no downtime).
  2. It’s safe to run whenever.
  3. No risk of data loss.
  4. AWS suggests that disk performance might be throttled during the modification, but we’ve never seen or heard about meaningful performance impact.
  5. In progress modifications in the past have prevented us from making backups, but this is something that’s really only impactful for large (> 500 GB) databases that take several hours to back up.