fix(destination-databricks): Add JDBC timeout configuration to prevent indefinite hangs (AI-Triage PR) (#74732)

Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: bot_apk <apk@cognition.ai>
Co-authored-by: Francis Genet <francis.genet@airbyte.io>

Author: 3 people

airbyte-integrations/connectors/destination-databricks/metadata.yaml

@@ -2,7 +2,7 @@ data:
   connectorSubtype: database
   connectorType: destination
   definitionId: 072d5540-f236-4294-ba7c-ade8fd918496
-  dockerImageTag: 3.3.7
+  dockerImageTag: 3.3.8
   dockerRepository: airbyte/destination-databricks
   githubIssueLabel: destination-databricks
   icon: databricks.svg

airbyte-integrations/connectors/destination-databricks/src/main/kotlin/io/airbyte/integrations/destination/databricks/DatabricksConnectorClientsFactory.kt

@@ -35,6 +35,17 @@ object DatabricksConnectorClientsFactory {
         return WorkspaceClient(config)
     }
 
+    // Socket timeout for JDBC read operations (seconds).
+    // Large MERGE / T+D queries can legitimately run for tens of minutes, so we set a generous
+    // ceiling. A value of 0 means "wait forever", which caused the 20+ hour hang in
+    // https://github.com/airbytehq/oncall/issues/11610.
+    private const val SOCKET_TIMEOUT_SECONDS = 3600
+
+    // Maximum time (seconds) the driver will retry connecting to a temporarily-unavailable
+    // warehouse (e.g. auto-paused). Default is 0 (no retry). 300 s (5 min) allows time for
+    // warehouse resume while still failing faster than the platform job deadline.
+    private const val TEMPORARILY_UNAVAILABLE_RETRY_TIMEOUT_SECONDS = 300
+
     fun createDataSource(config: DatabricksConnectorConfig): DataSource {
         val className = Driver::class.java.canonicalName
         Class.forName(className)
@@ -44,8 +55,18 @@ object DatabricksConnectorClientsFactory {
         // EnableArrow=0 flag is undocumented and disables ArrowBuf when reading data
         // Destinations only reads data for metadata or for comparison of actual data in tests. so
         // we don't need it to be optimized.
+        //
+        // SocketTimeout prevents indefinite hangs on long-running queries. See
+        // https://kb.databricks.com/dbsql/
+        //   job-timeout-when-connecting-to-a-sql-endpoint-over-jdbc
+        // TemporarilyUnavailableRetryTimeout caps the time the driver retries
+        // connecting to a paused/resuming warehouse.
         val jdbcUrl =
-            "jdbc:databricks://${config.hostname}:${config.port}/${config.database};transportMode=http;httpPath=${config.httpPath};EnableArrow=0"
+            "jdbc:databricks://${config.hostname}:${config.port}/${config.database}" +
+                ";transportMode=http;httpPath=${config.httpPath};EnableArrow=0" +
+                ";SocketTimeout=$SOCKET_TIMEOUT_SECONDS" +
+                ";TemporarilyUnavailableRetryTimeout=" +
+                "$TEMPORARILY_UNAVAILABLE_RETRY_TIMEOUT_SECONDS"
         when (config.authentication) {
             is BasicAuthentication -> {
                 datasource.setURL(

docs/integrations/destinations/databricks.md

@@ -96,6 +96,7 @@ This destination supports [namespaces](https://docs.airbyte.com/platform/using-a
 
 | Version | Date       | Pull Request                                                                                                        | Subject                                                                                                                                                                                |
 |:--------|:-----------|:--------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| 3.3.8   | 2026-03-11 | [74732](https://github.com/airbytehq/airbyte/pull/74732)                                                            | Add JDBC ConnectTimeout and SocketTimeout to prevent indefinite hangs when Databricks SQL warehouse is paused or unresponsive                                                          |
 | 3.3.7   | 2025-07-15 | [63311](https://github.com/airbytehq/airbyte/pull/63311)                                                            | Support arbitrary number of streams in findExisitngTable query                                                                                                                         |
 | 3.3.6   | 2025-03-24 | [56355](https://github.com/airbytehq/airbyte/pull/56355)                                                            | Upgrade to airbyte/java-connector-base:2.0.1 to be M4 compatible.                                                                                                                      |
 | 3.3.5   | 2025-03-07 | [55232](https://github.com/airbytehq/airbyte/pull/55232)                                                            | fix table name collision multiple connections same schema                                                                                                                              |