CodeBlanks Logo Icon CodeBlanks
Login Register

RPO and RTO

Learn recovery point and time objective concepts.

RPO and RTO

A disaster plan has two key targets:

  • RPO (Recovery Point Objective): the maximum acceptable period of data loss. A nightly backup makes the RPO 24 hours; an hourly backup 1 hour.
  • RTO (Recovery Time Objective): the longest acceptable time for the service to be working again.
              last backup        disaster                service back
   ───────────────┼──────────────────┼───────────────────────────┼──────►
                  └──── RPO ─────────┘└────────── RTO ────────────┘

Targets are chosen according to business needs: a payment system wants minutes, an internal wiki can live with hours.

Meeting the targets

To lower RPO: frequent backups, continuous WAL archiving in the database, or replication to another region. To lower RTO: fast rebuild with GitOps, a ready second cluster (warm standby), automatic DNS failover and regular drills.

service     RPO       RTO      method
odeme       1 min     15 min   database replication + warm standby
magaza      1 hour    1 hour   hourly Velero + GitOps
wiki        24 hours  1 day    nightly backup

Practice Hourly backup

To lower the RPO to 1 hour, write two commands: create a Velero schedule named saatlik that backs up magaza on the hour ("0 * * * *") and keeps it for 48 hours; then trigger a backup from the schedule right away (velero backup create --from-schedule saatlik).

Starter code

terminal
# buraya yaz

Sample solution

terminal
velero schedule create saatlik --schedule="0 * * * *" --include-namespaces magaza --ttl 48h
velero backup create --from-schedule saatlik

Expected output

Schedule "saatlik" created successfully.
Creating backup from schedule, all other filters are ignored.
Backup request "saatlik-20261103113005" submitted successfully.

Questions and answers

Review the lesson questions together with their answers. To test yourself, solve the lesson interactively.

1 If backups are taken every 6 hours, what is the worst-case RPO?

  • 1 hour
  • 24 hours
  • 0
  • 6 hours Correct answer

Explanation: If disaster strikes just before the next backup, the last 6 hours are lost.

2 The longest accepted time for the service to work again.

Question 
Target abbreviation: ____
Answer 
Target abbreviation: RTO

Explanation: RTO is time, RPO is data loss.

3 Match.

Term Meaning
RPO Data loss
RTO Downtime
warm standby Ready second cluster
WAL archive Continuous database backup

All lessons in Expert

1 Kubernetes internals

  1. What is etcd?
  2. kube-scheduler
  3. controller-manager
  4. kube-apiserver
  5. Reconcile loop
  6. Summary - Internals

2 Custom Resource Definition

  1. What is a CRD?
  2. CRD YAML
  3. Custom resource
  4. Spec and status
  5. Validation
  6. Summary - CRD

3 Operator pattern

  1. What is an Operator?
  2. Controller logic
  3. Operator SDK
  4. Custom resource management
  5. Example operator
  6. Summary - Operator

4 Service mesh - Istio

  1. What is a service mesh?
  2. What is Istio?
  3. Envoy sidecar
  4. Traffic management
  5. mTLS
  6. Summary - Service mesh

5 Multi-cluster

  1. What is multi-cluster?
  2. Cluster federation
  3. Kubeconfig
  4. Cross-cluster service
  5. Cluster registry
  6. Summary - Multi-cluster

6 OPA and Gatekeeper

  1. What is OPA?
  2. The Rego language
  3. What is Gatekeeper?
  4. ConstraintTemplate
  5. Constraint
  6. Summary - OPA

7 Disaster recovery

  1. Backup strategy
  2. etcd backup
  3. Velero
  4. Restore process
  5. RPO and RTO
  6. Summary - Disaster recovery

8 Expert final

  1. Challenge - CRD
  2. Challenge - Operator
  3. Challenge - Istio
  4. Challenge - OPA
  5. Challenge - Expert project