| Table of Contents | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
...
All systems and databases must be preserved for AKS Cluster and applications to be completely recoverable, utilizing a combination of namespace duplicates, incremental backups, and database backups.
The volatility of data determines the frequency of backups; the retention period is determined by the criticality of data, with a minimum retention period of 7 days.
It is essential to securely store at least one fully recoverable version off-site, in a different Azure region, or off-site storage.
Backup documentation should clearly outline all critical data and necessary support items for recovery. Restoration documentation should encompass procedures for both individual system/application failures and total data center disaster scenarios.
It is crucial to regularly review and update backup and recovery documentation to align with new technology and business changes.
Annual testing of recovery procedures is imperative to ensure effectiveness.
Backup Points
Backup Component | Backup Method | Backup Performed By | Recovery Region | Backup/Retention Frequency | Retention Period |
Database | Azure Database | Azure | Geo Redundant | Point In Time | 7 Days |
Adeptia Connect Namespace | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment such as PVC, ConfigMap, and Certificates. | Adeptia (Velero Scheduled Job) | Depend on Cluster | Every 8 hours | 14 Days |
Ingress | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment (such as PVC, ConfigMap, Certificate, etc.). | Adeptia (Velero Scheduled Job) | Depend on Cluster | Every 8 hours | 14 Days |
Logs (EFK) Namespace | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment (such as PVC, ConfigMap, Certificate, etc.). | Adeptia (Velero Scheduled Job) | Depend on Cluster | Every 8 hours | 14 Days |
Monitoring (Grafana) Namespace | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment (such as PVC, ConfigMap, Certificate, etc.). | Adeptia (Velero Scheduled Job) | Depend on Cluster | Every 8 hours | 14 Days |
Restore\Recovery Points
Recovery Use Case | Recovery Method | Performed By | Recovery Region | Restore Duration |
Namespace is accidentally lost | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment (such as PVC, ConfigMap, Certificate, etc.). | Adeptia | The recovery region will remain the same if the existing cluster is running smoothly. If a new cluster needs to be set up in a different region, the recovery process will be carried out in that new region. | 2-4 hours |
PVC Failed | Using Velero backup PVC object will be resorted | Adeptia | Recovery will remain in the same region if the existing cluster is operational. If a new cluster needs to be set up in a different region, recovery operations will be conducted in that new region. | 2-4 hours |
Kubernetes Failed | Connectria/Azure | 1-2 days(As per Connectria) | ||
Database/Server Failed or accidentally lost | Restore the database from the database backup and verify the connectivity of the application. | Adeptia/Connectria | 2-4 hours | |
Logs (EFK) Namespace | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment (such as PVC, ConfigMap, Certificate, etc.). | Adeptia | Recovery will remain in the same region if the existing cluster is operational. If a new cluster needs to be set up in a different region, recovery operations will be conducted in that new region. | 2-4 hours |
Monitoring (Grafana) Namespace | Using Velero backup, recover the entire namespace that includes components and objects of the application deployment (such as PVC, ConfigMap, Certificate, etc.). | Adeptia | Recovery will remain in the same region if the existing cluster is operational. If a new cluster needs to be set up in a different region, recovery operations will be conducted in that new region. | 2-4 hours |
...
Backups from customer environments follow this naming convention:
daily<<customername>>prod_YYYYMMDDHHmmSS
Object\Namespace Included
...