I know that HA is planned for render, but I do have a question about what happens if the service crashes between backups currently? Managed services usually offer continuous WAL backup as part of their backup offering or offer at least a persistent disk so the data is not lost. Losing one day of data is really not acceptable for production.
What is currently done by render exactly?
Hi @Sytten! Our managed postgres databases do use persistent disks under the hood. These disks have built-in redundancy so data loss due to a crash is very unlikely. The daily backups are provided less to guard against crashes and more to guard against unwanted writes/deletes (e.g. you accidentally drop a table or deploy a change that overwrites important data).
Ha thanks for the answer. I believe that should be added to the documentation.
As well as IOPS performance for those disks if possible.
Some people care both about having little ops and performance/reliability
Good point @Sytten! We’ll update our docs.