¶ Intro
This is automated RDS restore from backup (snapshot) process conntrolled by lambda function.
NOTE: that lambda invocation is idempotent. If it exeed 15min you can invoke again to continue
¶ Prerequisites
- make sure in your private AWS ECR you have repo rds-autorestore and image is available. If not you can refer to https://github.com/CTO2BPublic/rds-autorestore/blob/main/build.sh or ask CTO2B to push it
¶ Deploy
- select RDS Automatic Restore Lambda deployment from service catalog items
- update x3 items in deployment config
image_uri: <Your AWS account ID>.dkr.ecr.eu-central-1.amazonaws.com/rds-autorestore:0.0.6 #1
environment_variables:
SOURCE_DB: "dev-development-ab-bg-rds-mys" #2
SNAPSHOT_ID: "rds:dev-development-ab-bg-rds-mys-restored-2025-07-23-09-47" #3
¶ Restore process
- get in you aws console -> Lambda -> select your deployed lambda name
- select Test. Event JSON not matter
- start invocation by Test button
- once completed update your database revision number and press "Save" for refresh terraform state.
terraform:
spec:
persistentVolumeSize: 10Gi
keepLatestPodsOnly: true
variables:
rev: 6 # <- update number or add this if not exsist
multi_az: false
...
¶ Monitoring process
- use CloudWatch
- for logGroups search by lambda name and monitor the process
¶ Demo
