Apache APISIX: How to implement plugin orchestration in API Gateway wenming@apache.org Apache APISIX VP Ming Wen 温铭 • Co-founder @ api7.ai • VP and PMC member, Apache APISIX • committer, Apache creation • How to solve: GUI, Lego build • How children learn to program？ Scratch Plugin orchestration Demo How to implement？Step 1 • use DAG（Directed acyclic graph） for GUI • JSON of DAG Open source projects blooming：BFE、MOSN Low Code • PM can directly implement features by plugin orchestration • developers pay more attention to core of gateway, performance • Finally, the business and

ciphertext column query Orchestration Configuration Item Explanation # Omit data sharding, Read-Write split, and Data masking configuration. orchestration: name: orchestration_ds overwrite: true registry: registry: type: zookeeper namespace: orchestration serverLists: localhost:2181 ShardingSphere-3.x Data Sharding Configuration Item Explanation dataSources: ds0: !!org.apache.commons.dbcp.BasicDataSource ; Orchestration Configuration Item Explanation # Omit data sharding, Read-Write split configuration. orchestration: name: orchestration_ds overwrite: true registry: namespace: orchestration serverLists:

ShardingQueryAssistedEncryptor, it can help query encrypted data encryptor: # Encrypt name Orchestration Configuration Item Explanation dataSources: # Omit data source configurations shardingRule: read-write split rule configurations encryptRule: # Omit encrypt rule configurations orchestration: name: # Orchestration instance name overwrite: # Whether to overwrite local configurations with registry document, v5.1.1 DataSource dataSource = MasterSlaveDataSourceFactory.createDataSource(yamlFile); Orchestration Configuration Item Explanation dataSources: # Ignore data sources configuration shardingRule:

ShardingQueryAssistedEncryptor, it can help query encrypted data encryptor: # Encrypt name Orchestration Configuration Item Explanation dataSources: # Omit data source configurations shardingRule: read-write split rule configurations encryptRule: # Omit encrypt rule configurations orchestration: name: # Orchestration instance name overwrite: # Whether to overwrite local configurations with registry createDataSource(yamlFile); 7.9. API Change Histories 380 Apache ShardingSphere document, v5.1.2 Orchestration Configuration Item Explanation dataSources: # Ignore data sources configuration shardingRule:

spring.shardingsphere.orchestration.name= # 治理实例名称 spring.shardingsphere.orchestration.overwrite= # 本地配置是否覆盖注册中心配置。如果可覆盖， 每次启动都以本地配置为准 spring.shardingsphere.orchestration.registry.type= # 配置中心类型。如：zookeeper 配置中心类型。如：zookeeper spring.shardingsphere.orchestration.registry.server-lists= # 连接注册中心服务器的列表。 包括 IP 地址和端口号。多个地址用逗号分隔。如: host1:2181,host2:2181 spring.shardingsphere.orchestration.registry.namespace= # 注册中心的命名空间 注册中心的命名空间 spring.shardingsphere.orchestration.registry.digest= # 连接注册中心的权限令牌。缺省为不需 要权限验证 spring.shardingsphere.orchestration.registry.operation-timeout-milliseconds= # 操作 超时的毫秒数，默认 500 毫秒 spring.shardingsphere

省略数据源配置 shardingRule: # 省略分片规则配置 masterSlaveRule: # 省略读写分离规则配置 encryptRule: # 省略数据脱敏规则配置 orchestration: name: # 治理实例名称 overwrite: # 本地配置是否覆盖注册中心配置。如果可覆盖，每次启动都以本地配置为准 registry: # 注册中心配置 type: # 必须配置的项目为 schemaName,dataSources, 并且 shardingRule, masterSlaveRule, 配置其中一个 (注意, 除非 server.yaml 中定义了 Orchestration, 否则必须至少 有一个 config-xxxx 配置文件), 除此之外的其他项目为可选项 schemaName: test # schema 名称, 每个文件都是单独的 schema 治理 配置项说明 dataSources: # 省略数据源配置 shardingRule: # 省略分片规则配置 masterSlaveRule: # 省略读写分离规则配置 orchestration: name: # 数据治理实例名称 overwrite: # 本地配置是否覆盖注册中心配置。如果可覆盖，每次启动都以本地配置为准 7.9. API 变更历史 326 Apache

省略数据源配置 shardingRule: # 省略分片规则配置 masterSlaveRule: # 省略读写分离规则配置 encryptRule: # 省略数据脱敏规则配置 orchestration: name: # 治理实例名称 overwrite: # 本地配置是否覆盖注册中心配置。如果可覆盖，每次启动都以本地配置为准 registry: # 注册中心配置 type: # 必须配置的项目为 schemaName,dataSources, 并且 shardingRule, masterSlaveRule, 配置其中一个 (注意, 除非 server.yaml 中定义了 Orchestration, 否则必须至少 有一个 config-xxxx 配置文件), 除此之外的其他项目为可选项 schemaName: test # schema 名称, 每个文件都是单独的 schema 治理 配置项说明 dataSources: # 省略数据源配置 shardingRule: # 省略分片规则配置 masterSlaveRule: # 省略读写分离规则配置 orchestration: name: # 数据治理实例名称 overwrite: # 本地配置是否覆盖注册中心配置。如果可覆盖，每次启动都以本地配置为准 7.9. API 变更历史 322 Apache

0-beta Orchestration Configuration Item Explanation #Omit data source, data sharding, readwrite split and data masking configurations spring.shardingsphere.orchestration.name= #Orchestration instance instance name spring.shardingsphere.orchestration.overwrite= #Whether to overwrite local configurations with registry center configurations; if it can, each initialization should refer to local configurations configurations spring.shardingsphere.orchestration.registry.type= #Registry center type. Example:zookeeper spring.shardingsphere.orchestration.registry.server-lists= #The list of servers that connect to registry

省略数据源配置 shardingRule: # 省略分片规则配置 masterSlaveRule: # 省略读写分离规则配置 encryptRule: # 省略数据脱敏规则配置 orchestration: name: # 治理实例名称 overwrite: # 本地配置是否覆盖注册中心配置。如果可覆盖，每次启动都以本地配置为准 registry: # 注册中心配置 type: # 必须配置的项目为 schemaName,dataSources, 并且 shardingRule, masterSlaveRule, 配置其中一个 (注意, 除非 server.yaml 中定义了 Orchestration, 否则必须至少 有一个 config-xxxx 配置文件), 除此之外的其他项目为可选项 schemaName: test # schema 名称, 每个文件都是单独的 schema 治理 配置项说明 dataSources: # 省略数据源配置 shardingRule: # 省略分片规则配置 masterSlaveRule: # 省略读写分离规则配置 orchestration: name: # 数据治理实例名称 overwrite: # 本地配置是否覆盖注册中心配置。如果可覆盖，每次启动都以本地配置为准 7.9. API 变更历史 338 Apache

frameworks for identifying promising use cases, clear patterns for designing agent logic and orchestration, and best practices to ensure your agents run safely, predictably, 
 and effectively. After tradeoffs related to task complexity, latency, and cost. As we’ll see in the next section on Orchestration, you might want to consider using a variety 
 of models for different tasks in the workflow. hand-off a customer service ticket to a human. Orchestration Agents themselves can serve as tools for other agents—see the Manager Pattern in the Orchestration section. Refund agent, Research agent, Writing