Skip to main content

Set up Read Replica

A Read Replica is a read-only copy of your production InsuranceSuite application database. You can use SQL tools to query current production data without querying the primary database directly.

Note:

Read Replica access is an add-on for Guidewire Cloud Platform. To enable this feature, contact Guidewire.

Performance considerations for high-volume queries​

Read Replica doesn't eliminate performance risk. Large, complex, or long-running queries, especially full-table extracts and recurring batch workloads, can consume reader resources and might affect the writer nodes and application performance.

For high-volume data extraction, reporting, data warehousing, and recurring ETL or downstream integration workloads, use Cloud Data Access (CDA) instead. CDA performs an initial bulk load and then streams incrementally changing data in near real time to an Amazon S3 bucket in Apache Parquet format. You can download the data into your own data lake, warehouse, or processing environment and run extraction and transformation workloads there.

Supported use cases​

Use a Read Replica for targeted workloads such as:

  • Small-volume, short-running queries against current InsuranceSuite data.
  • Operational lookups that require direct SQL access and the latest available database state.
  • Development, troubleshooting, or data-validation activities where the query scope is controlled.

Use CDA instead of a Read Replica for:

  • Initial or full-table data extraction.
  • Recurring ETL or batch extraction jobs.
  • Data warehouse and data lake loading.
  • High-volume reporting or analytics.
  • Downstream integrations that require incremental change data or historical change tracking.

You can use your own SQL tools to query production data. Queries can return unmasked personally identifiable information (PII). Handle PII securely and follow data protection policies from your organization.

Read Replica doesn't support change data capture (CDC). If you need CDC capability, consider using Guidewire CDA.

Set up real-time data access​

Real-time data queries use an AWS PrivateLink connection for performance and security. For details about PrivateLink, see Network connectivity with Cloud Platform.

Read Replica adds a reader instance to the InsuranceSuite database cluster. After you set up your network, Guidewire provides a PrivateLink reader endpoint. Use the endpoint to query the database.

Additional reader instances in a Read Replica architecture.

Guidewire provides a user account for Read Replica access. You can request up to three user accounts for each InsuranceSuite application on each planet. To request additional accounts, submit a case in Guidewire Community using the Modify Read Replica (DB-0011) request template.

Read Replica usage guidelines​

To improve performance and ensure that you don't overload the server, follow these guidelines when constructing your queries:

  • Use efficient queries

    Avoid using SELECT * and always include a WHERE clause to filter results.

  • Avoid idle connection timeouts

    For enhanced security, Guidewire Cloud Platform enforces an idle network connection timeout of 20 minutes. A Read Replica query times out if the database reader node takes longer than 20 minutes to start responding. The timeout doesn't apply while the reader actively sends a response.

    If you encounter timeout issues, optimize your queries.

  • Reduce query complexity

    A long-running query can affect the writer nodes and reduce application performance. Optimize your queries to avoid this impact. By default, Read Replica queries time out after two hours, and the system terminates queries that exceed this limit.

    If you require queries to run longer than two hours, contact Guidewire Support to disable the Read Replica query timeout.