In order to be able to quickly restore from backups as an SRE I want a tested, documented process for restoring influxdb from our s3 backups.
Acceptance Criteria
[ ] Operations docs are updated with a tested process for restoring data into our existing influxdb deployment from s3 backups
[ ] Operations docs are updated with a tested process for restoring data into a newly created influxdb deployment running along side our existing deployment
Implementation Sketch
I would like to see us test two scenarios that would require restoring from backup.
Normal recovery
We replace the data in our existing influxdb deployment with the data from our backups.
IR recovery
We can stand up up a new influxdb deployment next to our existing influxdb deployment with data loaded from a specific point-in-time.
If our influxdb system is breached we may need to preserve it in it's current state for forensic purposes and would need to spin up a new influxdb deployment concurrently to compare data from known-good-backups to what is in the compromised deployment.
In order to be able to quickly restore from backups as an SRE I want a tested, documented process for restoring influxdb from our s3 backups.
Acceptance Criteria
Implementation Sketch
I would like to see us test two scenarios that would require restoring from backup.
Normal recovery
We replace the data in our existing influxdb deployment with the data from our backups.
IR recovery
We can stand up up a new influxdb deployment next to our existing influxdb deployment with data loaded from a specific point-in-time.
If our influxdb system is breached we may need to preserve it in it's current state for forensic purposes and would need to spin up a new influxdb deployment concurrently to compare data from known-good-backups to what is in the compromised deployment.