Data backup, data restoration, and data durability for Pega Cloud

This article is part of the Pega Cloud Subscription Documentation.

Pega Cloud production services provides various levels of backup and redundancy for client data, to provide full recovery of environments in the event of service disruption or failure. 

Pega’s approach involves the concepts of both data durability and data restoration.  Data durability defines our ability to keep data stored in Cloud Data Storage and Decision Data Storage consistent, while heavily mitigating the possibility of data loss when there is a failure within the data stores.  Achieving data durability in this scenario is part of our native high-availability architecture.  Our data restoration practice is leveraged when the service must be recovered from backup, because of an unlikely and catastrophic failure within a region. 

• RPO is the maximum amount of time during which data might be lost from Cloud Data Storage or Decision Data Storage. 
• RTO is the target time to restore the Pega Cloud Environment(s)  

Data durability

Data durability provides consistency and resiliency of data when there is a failure of active resources, and high-availability processes engage to recover these critical systems.  These high availability concepts include:  

1. When a primary database is impacted, and the standby database must take over operation as the primary. 

2. When a node in a cluster is impacted, and the service must recover the failed node. 

The Pega Cloud datastore that provides Cloud Data Storage uses replication services that is designed to protect data by deploying a real-time standby in a different zone and replicating the data synchronously between primary and standby.  Synchronous replication means that for a transaction to be viewed as successfully committed by Pega, it must be written to both the primary and standby database.  This means that committed transactions are durably written to multiple zones, heavily mitigating the potential for data loss.   

Decision Data Storage leverages a multi-node clustered architecture spanning multiple zones, and a replication strategy so that a node or a zone failure does not impact the durability of data. 

In the event of these failures, Pega provides a recovery point objective (RPO) of approximately 1 minute.  In many scenarios, because of the strength of the replication strategy being employed, this value will be less.  Automated failover and recovery of the service has a recovery time objective (RTO) of approximately 4 minutes.

Data restoration from backup

A Pega Cloud subscription includes additional data management services, such as a data backup and restore service that continuously backs up the client’s Cloud Data Storage data for each Environment (up to 30 days in Pega Cloud AWS regions, and up to 7 days in Pega Cloud GCP regions).

The Pega Cloud data backup services back up client Cloud Data Storage in all production Environmentson an ongoing basis with an RPO of no more than 10 minutes in Pega Cloud AWS regions, and 24 hours in Pega Cloud GCP regions.  For Decision Data Storage, the RPO is 10 minutes for all AWS and GCP regions.  This is the RPO during a scenario where both the primary and standby database or the cluster have suffered an unrecoverable failure, and the service must be restored in full, from backup.  These are rare catastrophic events where the capabilities in place to provide data durability described above are not viable.  The time required (RTO) for a database backup restoration depends on the size of the deployed Environments. 

Pega Cloud reserves the right to update these restoration projections based on improvements to existing services, new services that Pega Cloud might offer, or any changes in the support of data or regular disaster recovery testing.

The number of monthly data restorations Pega Cloud provides depends on the supported Environment type: 

• Production Environments can receive unlimited, full data restoration. 
• Non-production Environments can receive up to three full database restorations per month. 

Pega Cloud restores complete backups only.  Pega recommends that a restoration copies backup data to the Environment type of its source. For example, Pega recommends that a backup of a production Environment be restored to a production Environment. 

If the target Environment requires more storage to support the amount of data being restored, it is the client's responsibility to purchase the additional storage needed.

To request that a Pega Cloud database be restored from a prior backup, the client should open a service request through My Support Portal.

Note:  Legacy Web Chat (formerly Pega Chat), Co-Browse, Digital Messaging, Voice AI, and Workforce Intelligence production Environments all provide ongoing backup for client data, to provide full restoration of Environments in the event of service disruption or failure.