1.1.1.1 Release 26ai - February 2026
This section lists the new features and upgrades in GGSA 26ai.
1.1.1.1.1 GGSA URL
Use the following URLs to connect to GGSA 26ai:
For secure connection, use:
https://<url>:<port>/osa/index.html.
For non-secure connection, use:
http://<url>:<port>/osa/index.html.
By default GGSA will start in SSL enabled mode, if values are not
provided in the ssl.conf file, then the start
script will generate a self-signed certificate for you.
-
To disable SSL mode set NEED_SSL=false in the ssl.conf file.
-
The default port is 9080, when
SSL=False.
1.1.1.1.2 Spark and Kafka Versions Upgrade
- Spark library version is upgraded to 4.0.1
- Kafka library version is upgraded to 4.1.0
For more information on downloading the libraries, and installing GGSA, see Installing GoldenGate Stream Analytics.
1.1.1.1.3 Similarity Search Pattern
The Similarity Search pattern is added in the GGSA release 26ai, to enable vector-based search in the Oracle database tables. For more information about using this pattern, see Applying Vector-Based Similarity Search.
1.1.1.1.4 FromJson Pattern
FromJson pattern flattens nested JSON attributes into new columns while preserving all original input fields. Arrays are preserved as JSON strings (no row expansion).
Use this pattern to convert nested JSON into simple columns.
For more information about creating and using this pattern, see Transforming Nested JSON into Simple Columns.
1.1.1.1.5 Oracle Generative AI Agent Pattern
The Generative AI Agent pattern integrates an Oracle Cloud Generative AI service call directly into your Stream Explorer pipeline.
You can use this new pattern for:
Natural-language enrichment - To add summaries, sentiment, explanations, or decisions to every event.
No-code configuration - Just select a connection, paste the Agent OCID, write a prompt.
Full schema continuity - All upstream attributes stay intact; nothing is lost.
Prompt transparency - Both the generated prompt and the agent’s reply are returned for auditing dashboards.
For more information about using this pattern, see Oracle Generative AI Agent Pattern.
1.1.1.1.6 Vector Database as a Target
You can create a vector database target to store the data as a target reference, to enable a similarity search on the incoming data. For more information on implementing vector database as a target, see Creating a Database Target.
1.1.1.1.7 GoldenGate Data Stream Integration
Oracle GoldenGate Data Stream is integrated with GGSA, enabling the ingestion and processing of real-time change data capture (CDC) events, over a secure WebSocket connection (data-stream).The integration streamlines the ingestion of transactional data by leveraging GoldenGate 26ai's new Data Stream feature, eliminating the traditional Distribution Paths and Replicats. The Data Stream will feed CDC changes directly into Spark applications deployed through GGSA.
For more information about creating and using GoldenGate Data Stream, see Creating a GoldenGate Data Stream.
1.1.1.1.8 "$" Field Path in the Target
As part of the GG DataStream integration feature in GGSA 26ai, the "$" field path in the target is supported, as an equivalent to the similar field path in a source stream.
This feature allows you to write raw JSON string in target. "$" field path is supported for targets which support JSON as content type and the UI supports "$" value in the Field Path.
You can specify "$" as a field path in a manual shape stream.
1.1.1.1.9 JSONPath Function
The JSONPath function extracts the column name value from the given JSON string. The column name has to be in the format, $.columnname.
jsonpath(column name holding the jsonstring, key attribute path to be extracted) -
parses and extracts the key attributes value from the given JSON column string. The
key path has to be in the format
jsonColumnName_KeyAttributePath.
For more information about using the JSONPath function, see Using JSONPath Functions.
1.1.1.1.10 Embedded Ignite Cache, GGBD and Ignite Clusters on OCI GGSA Desupported
From GGSA 26ai, Embedded Ignite Cache, GGBD and Ignite Clusters on OCI GGSA are desupported.
1.1.1.1.11 Embedded Coherence Cache
GGSA implements an Embedded Coherence Cache. Coherence is only used with cache-enabled DB reference and full records. In case of GoldenGate, the pipeline persists incoming events containing full record, and also updates the cache with modified events, making available the latest modified records.
For more information about managing and monitoring embedded coherence cache, see Embedded Coherence Cache.