How does Panther handle high availability?
The backups are stored in the same region, but separate from the Panther application (i.e. in an S3 bucket not controlled by the panther application, and that the panther application can’t interact with besides to backup logs there). In theory even if the Panther application explodes and tries to destroy all the data it has, it won’t affect the backups.
We primarily backup customer log data, not customer configuration data. Those backups last at least a year in S3. Customer configuration data has 30 day point in time recovery from dynamodb so we can recover deleted data for up to 30 days.The log data backup is separate from Snowflake, so in theory of snowflake blew up we’d still have 1 year of customer log data. DynamoDB and S3 both have lots of high availability promises from AWS. I’m sure we can find some docs on those if they want.
Customers who have a BI user created in Snowflake, and Snowflake Connected App deployments can still log in and view their data tables in Snowflake directly.