Back to Details

sf-datacloud-connect

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

sf-datacloud-connect

sf-datacloud-connect

Use this skill for the connection plane.
Connector catalog: references/connector-implementation-cards.md
Beast references:
  • Beast preflight: docs/beast-preflight.md
  • Phase proof matrix: docs/phase-proof-matrix.json
  • Public operating model: docs/operating-model.md
  • Interoperability decision map: docs/data360/interoperability-decision-map.md
  • Developer Guide index: docs/data360/developer/index.md
  • Proof ledger: docs/proof-ledger.md
  • Public LLM map: docs/llms.txt
  • Limits source precedence: docs/data360/limits-source-precedence.md
  • For exact Salesforce behavior, fetch official Help/Developer docs on demand with 
    sf-docs
    .
  • For endpoint shape, use OpenAPI from the official spec or the user-supplied Swagger before writing payloads.
此skill适用于连接层。
连接器目录：references/connector-implementation-cards.md
相关参考文档：
  • Beast预检查：docs/beast-preflight.md -阶段验证矩阵：docs/phase-proof-matrix.json -公开运营模型：docs/operating-model.md -互操作性决策图：docs/data360/interoperability-decision-map.md -开发者指南索引：docs/data360/developer/index.md -验证记录：docs/proof-ledger.md -公开LLM映射：docs/llms.txt -限制优先级来源：docs/data360/limits-source-precedence.md -若需了解Salesforce的准确行为，可通过
    sf-docs
    按需获取官方帮助/开发者文档。 -编写请求体前，优先使用官方规范中的OpenAPI或用户提供的Swagger来确定端点结构。

Prefer these Connect API families

优先选择以下Connect API系列

  • GET /ssot/connectors
  • GET /ssot/connectors/:connectorType
  • GET /ssot/connections
  • POST /ssot/connections
  • GET /ssot/connections/:connectionId
  • PATCH /ssot/connections/:connectionId
  • POST /ssot/connections/actions/test
  • POST /ssot/connections/:connectionId/actions/test
  • GET/PUT /ssot/connections/:connectionId/schema
  • GET /ssot/connectors
  • GET /ssot/connectors/:connectorType
  • GET /ssot/connections
  • POST /ssot/connections
  • GET /ssot/connections/:connectionId
  • PATCH /ssot/connections/:connectionId
  • POST /ssot/connections/actions/test
  • POST /ssot/connections/:connectionId/actions/test
  • GET/PUT /ssot/connections/:connectionId/schema

Default workflow

默认工作流

  1. inspect connector metadata
  2. classify the connector as ingestion, query federation, file federation, sharing, or hybrid
  3. inspect existing connections
  4. test before creating when possible
  5. upload or verify schema for ingestion-style connectors
  6. hand off to Prepare or Retrieve once the connection and pattern are clear
  1. 检查连接器元数据
  2. 将连接器分类为ingestion、查询联邦、文件联邦、共享或混合类型
  3. 检查现有连接
  4. 尽可能在创建前进行测试
  5. 为ingestion类型的连接器上传或验证schema
  6. 在连接和模式明确后，移交至Prepare或Retrieve环节

Rules

规则

  • Prefer REST payload inspection over guessing connector parameters.
  • Reuse the OpenAPI catalog through sf-datacloud-connectapi; Postman is optional comparison only.
  • For local development, CLI-auth or direct access token is usually faster than creating a new connected app.
  • Review source prerequisites before creating streams. Help docs separate source configuration from stream setup.
  • For connector-count and CRM-org-count questions, use current Data 360 Limits and Guidelines first. Follow Data Services Billable Usage Types when the current page points there. Use legacy Customer Data Platform CRM-org limits only when CDP is explicitly in scope or as a labeled comparison.
  • Inspect connector metadata and test the connection before handing off to Prepare.
  • Treat a successful connection test as control-plane proof only. Before calling a source ready, prove source-object discovery, source-side grants, and any connector-specific eligibility rules; for file federation, include table kind/format, catalog access, and the underlying storage path.
  • When the requested scope continues past connection setup, use the readiness chain 
    connection -> source discovery -> stream/DLO -> DLO query -> DMO mapping -> DMO query
    . Report downstream links as unproved instead of inferring them from healthy authentication.
  • If data spaces are involved, confirm where the connection, stream, and resulting DLOs are scoped.
  • For external lakehouses, choose the interoperability pattern before creating assets: ingestion for canonical governance, live query for maximum freshness, accelerated query for frequent reads with stale tolerance, file federation for large object-store/open-table workloads, or hybrid for governed core plus fresh edge.
  • Capture source-system cost and governance assumptions for federated connections. Query federation can depend on external compute and source-side policies; file federation depends on storage access, table format, partitioning, and Data 360 compute.
  • For Databricks zero-copy work, classify the exact connector mode before giving network guidance: query federation, accelerated query, file federation, data share, or batch ingestion. Do not let the umbrella "Databricks connector" hide different runtime paths.
  • Separate 
    Salesforce Private Connect
    from 
    Private Connect for Data 360
    . Use current official docs and tenant validation before claiming Databricks private routing, especially for AWS-hosted Databricks. Public docs have been explicit for Databricks-on-Azure PNR setup, while AWS claims need exact source support or live org proof.
  • For Databricks query federation, treat Data 360 IP allowlisting as both a setup gate and a runtime dependency for query, acceleration refresh, mapping, and downstream workloads. The primary allowlist point is Databricks-side workspace or SQL warehouse network control; customer-managed AWS firewalls, proxies, or PrivateLink-style layers add separate enforcement points.
  • For Databricks file federation, map both network legs: Data 360 to the Databricks or Unity Catalog endpoint, and Data 360 to the underlying storage layer such as S3. Recheck current docs before promising PrivateLink support; prior official docs required public accessibility for Unity Catalog and storage and did not support AWS PrivateLink or Azure Private Link for this connector mode.
  • When connector metadata is available, inspect it before payload design. In a prior Databricks query-federation surface, useful fields included 
    hasPrivateNetworkRoute
    , 
    outboundnetworkconnection
    , 
    jdbc_connection_url
    , and 
    httpPath
    ; for Databricks file federation, inspect catalog endpoint, storage type, storage credentials, and identity-provider fields.
  • Classify integration path before implementation: built-in Salesforce connector, external connector, Ingestion API bulk/streaming, Amazon S3 data stream, Salesforce Interactions SDK for web behavior, Engagement Mobile SDK for mobile events, MuleSoft, or zero-copy/federation.
  • For API-driven ingestion, confirm OAuth scope requirements and whether the work is bulk historical load, recurring bulk load, or small-batch streaming.
  • 优先检查REST请求体，而非猜测连接器参数。
  • 通过sf-datacloud-connectapi复用OpenAPI目录；Postman仅作为可选对比工具。
  • 本地开发时，CLI认证或直接访问令牌通常比创建新的连接应用更快。
  • 创建流之前，查看源系统的先决条件。帮助文档会将源配置与流设置分开说明。
  • 对于连接器数量和CRM组织数量的问题，优先使用当前的Data 360限制与指南。若当前页面指向Data Services计费使用类型，则遵循该类型。仅当CDP明确在范围内或作为标注对比时，才使用旧版客户数据平台CRM组织限制。
  • 在移交至Prepare环节前，检查连接器元数据并测试连接。
  • 仅将成功的连接测试视为控制层验证。在判定源系统就绪前，需验证源对象发现、源端权限以及任何连接器特定的资格规则；对于文件联邦，需包含表类型/格式、目录访问权限以及底层存储路径。
  • 当请求范围超出连接设置时，使用就绪链
    connection -> source discovery -> stream/DLO -> DLO query -> DMO mapping -> DMO query
    。对于下游环节，仅报告未验证状态，而非从正常认证中推断其可用性。
  • 若涉及数据空间，确认连接、流及生成的DLO的作用域。
  • 对于外部湖仓，在创建资产前选择互操作性模式：ingestion用于规范治理，实时查询用于保证最大新鲜度，加速查询用于频繁读取且可容忍数据过时的场景，文件联邦用于大型对象存储/开放表工作负载，混合模式则用于受控核心加新鲜边缘的场景。
  • 记录联邦连接的源系统成本与治理假设。查询联邦依赖外部计算和源端策略；文件联邦依赖存储访问权限、表格式、分区以及Data 360计算资源。
  • 对于Databricks零拷贝工作，在提供网络指导前明确连接器的具体模式：查询联邦、加速查询、文件联邦、数据共享或批量ingestion。不要让“Databricks连接器”这个统称掩盖不同的运行路径。
  • 区分
    Salesforce Private Connect
    与
    Private Connect for Data 360
    。在声称Databricks私有路由可用前，使用当前官方文档和租户验证，尤其是AWS托管的Databricks。官方文档对于Azure上的Databricks PNR设置有明确说明，而AWS相关的声明需要确切的源系统支持或实际组织验证。
  • 对于Databricks查询联邦，将Data 360 IP白名单视为设置关卡和运行时依赖，适用于查询、加速刷新、映射及下游工作负载。主要的白名单设置点是Databricks端的工作区或SQL仓库网络控制；客户管理的AWS防火墙、代理或PrivateLink类层会增加额外的验证点。
  • 对于Databricks文件联邦，映射两条网络链路：Data 360到Databricks或Unity Catalog端点，以及Data 360到底层存储层（如S3）。在承诺支持PrivateLink前，重新查看当前文档；此前的官方文档要求Unity Catalog和存储具备公共访问性，且该连接器模式不支持AWS PrivateLink或Azure Private Link。
  • 当连接器元数据可用时，在设计请求体前先检查它。在之前的Databricks查询联邦界面中，有用的字段包括
    hasPrivateNetworkRoute
    、
    outboundnetworkconnection
    、
    jdbc_connection_url
    和
    httpPath
    ；对于Databricks文件联邦，检查目录端点、存储类型、存储凭证和身份提供者字段。
  • 在实施前分类集成路径：Salesforce内置连接器、外部连接器、Ingestion API批量/流式、Amazon S3数据流、用于Web行为的Salesforce Interactions SDK、用于移动端事件的Engagement Mobile SDK、MuleSoft或零拷贝/联邦。
  • 对于API驱动的ingestion，确认OAuth范围要求，以及工作类型是批量历史加载、定期批量加载还是小批量流式加载。

Validation gate

验证关卡

Connection work is not done until connector metadata is understood, auth is healthy, an eligible source object and its grants are proven, schema is discovered or uploaded, the integration pattern is explicit, and the next Data Stream/DLO/federated-query step is clear.
连接工作完成的标准是：理解连接器元数据、认证正常、验证合格的源对象及其权限、发现或上传schema、明确集成模式，且清楚后续数据流/DLO/联邦查询步骤。

Doc-Synced Notes

文档同步说明

<!-- SF_DOC_SYNC_START:connect-data-help-side -->
<!-- SF_DOC_SYNC_START:connect-data-help-side -->

Connect Data: Help-Side Concepts (Data Sources, Data Streams, Schedules)

连接数据：帮助端概念（数据源、数据流、调度）

Distilled from official Salesforce sources only.
Sources:
  • data.c360_a_connectors.htm — Data Sources in Data 360
  • data.c360_a_connection_tasks.htm — Data Source Configuration in Data 360
  • data.c360_a_data_streams.htm — Data Streams in Data 360
  • data.c360_a_data_stream_schedule.htm — Data Stream Schedule
  • data.c360_a_data_streams_tab.htm — Data Streams Tab Navigation
  • developer.salesforce.com/docs/data/data-cloud-int/guide/c360-a-create-ingestion-data-stream.html — Create an Ingestion API Data Stream
  • developer.salesforce.com/docs/data/data-cloud-int/guide/c360-a-create-sftp-data-stream.html — Create an SFTP Data Stream
Data Source vs Data Stream:
  • A Data Source is the connector-level connection (auth, endpoint, credentials). Configured once per source system.
  • A Data Stream is an ingestion pipeline from a Data Source into a specific Data Lake Object. Many streams can use one source.
  • The "Data Sources" tab manages connections; the "Data Streams" tab manages ingestion pipelines and schedules.
Creating a Data Stream (UI flow):
  1. App Launcher → Data Streams → New (or Data Streams tab → New).
  2. Choose connector source (Ingestion API, SFTP, Salesforce CRM, S3, Snowflake, etc.).
  3. Select or create the Data Lake Object (DLO) — the landing object in the Data 360 data lake.
  4. Assign a label, API name, and category (Profile, Engagement, Other).
  5. Choose Primary Key — required, must uniquely identify each record.
    • Platform Events: Event Identifier field.
    • Ingestion API / SFTP: any unique field.
  6. Map fields if needed (or use auto-detected schema).
  7. Choose Refresh Mode: Incremental (insert new), Full Refresh, or Partial (Profile/Other only — partial record updates without full replace).
  8. Configure Schedule (see below).
  9. Click Deploy.
  10. Optionally click Refresh Data Stream Immediately to start ingest right after deployment.
Refresh modes:
ModeBehaviorBest for
IncrementalInserts/updates new records by primary keyAppend-only or high-volume streams
Full RefreshReplaces all records on every runReference data, small dimension tables
PartialUpdates specified fields without full replaceProfile/Other category, partial CDC
Schedule options:
  • Frequencies: hourly (where supported), daily, weekly, monthly, or manual-only.
  • Different connectors support different schedule granularities; check connector page.
  • Manual refresh always available — overrides schedule when invoked.
  • Concurrent refresh limits apply per org; long-running streams may defer if capacity is constrained.
Data Streams Tab Navigation:
  • Lists all streams with status (Active, Error, Inactive, Processing).
  • Filter by data source, refresh mode, or DLO.
  • Inspect stream history (run logs, record counts, error counts).
  • Edit, deactivate, or refresh from this tab.
Pre-flight checks before creating a stream:
  • Connector authenticated and tested? (
    POST /ssot/connections/actions/test
    )
  • Schema discovered or uploaded? (Ingestion API requires OAS/YAML upload.)
  • DLO category chosen (Profile, Engagement, Other)?
  • Primary key chosen?
  • Refresh mode aligned with use case?
  • Schedule frequency aligned with downstream consumer cadence?
  • Data space scoping correct?
Common errors and remedies:
  • "Schema mismatch" → re-upload schema or align field types in the source.
  • "Primary key not unique" → choose a different field or composite key not supported (must use single-field uniqueness).
  • "Authentication expired" → rotate credentials in the Data Source, then retest the connection.
  • Stream stuck in Processing → check connector page for source-specific rate limits; verify source-side scheduling.
<!-- SF_DOC_SYNC_END:connect-data-help-side --> <!-- SF_DOC_SYNC_START:developer-integration-catalog -->
仅提炼自Salesforce官方来源。
来源文档：
  • data.c360_a_connectors.htm — Data 360中的数据源
  • data.c360_a_connection_tasks.htm — Data 360中的数据源配置
  • data.c360_a_data_streams.htm — Data 360中的数据流
  • data.c360_a_data_stream_schedule.htm — 数据流调度
  • data.c360_a_data_streams_tab.htm — 数据流标签页导航
  • developer.salesforce.com/docs/data/data-cloud-int/guide/c360-a-create-ingestion-data-stream.html — 创建Ingestion API数据流
  • developer.salesforce.com/docs/data/data-cloud-int/guide/c360-a-create-sftp-data-stream.html — 创建SFTP数据流
数据源 vs 数据流：
  • 数据源是连接器级别的连接（认证、端点、凭证）。每个源系统只需配置一次。
  • 数据流是从数据源到特定Data Lake Object的摄入管道。一个数据源可对应多个数据流。
  • “数据源”标签页管理连接；“数据流”标签页管理摄入管道和调度。
创建数据流（UI流程）：
  1. 应用启动器 → 数据流 → 新建（或数据流标签页 → 新建）。
  2. 选择连接器源（Ingestion API、SFTP、Salesforce CRM、S3、Snowflake等）。
  3. 选择或创建Data Lake Object (DLO) — Data 360数据湖中的落地对象。
  4. 分配标签、API名称和类别（Profile、Engagement、Other）。
  5. 选择主键 — 必填，必须能唯一标识每条记录。
    • 平台事件：事件标识符字段。
    • Ingestion API / SFTP：任意唯一字段。
  6. 如有需要，映射字段（或使用自动检测的schema）。
  7. 选择刷新模式：增量（插入新记录）、全量刷新或部分刷新（仅Profile/Other类别支持 — 无需全量替换即可更新指定字段）。
  8. 配置调度（见下文）。
  9. 点击部署。
  10. 可选择点击立即刷新数据流，在部署后立即开始摄入。
刷新模式：
模式行为适用场景
增量通过主键插入/更新新记录仅追加或高流量数据流
全量刷新每次运行替换所有记录参考数据、小型维度表
部分刷新更新指定字段，无需全量替换Profile/Other类别、部分CDC场景
调度选项：
  • 频率：每小时（支持的连接器）、每日、每周、每月或仅手动。
  • 不同连接器支持不同的调度粒度；请查看连接器页面。
  • 始终支持手动刷新 — 调用时会覆盖调度。
  • 每个组织有并发刷新限制；长时间运行的流可能因容量不足而延迟。
数据流标签页导航：
  • 列出所有流及其状态（活跃、错误、 inactive、处理中）。
  • 可按数据源、刷新模式或DLO筛选。
  • 查看流历史（运行日志、记录数、错误数）。
  • 可在此标签页编辑、停用或刷新流。
创建流前的预检查：
  • 连接器已认证并测试通过？(
    POST /ssot/connections/actions/test
    )
  • schema已发现或上传？（Ingestion API需要上传OAS/YAML。）
  • 是否已选择DLO类别（Profile、Engagement、Other）？
  • 是否已选择主键？
  • 刷新模式是否与用例匹配？
  • 调度频率是否与下游消费者节奏匹配？
  • 数据空间作用域是否正确？
常见错误及解决方法：
  • “Schema不匹配” → 重新上传schema或对齐源端字段类型。
  • “主键不唯一” → 选择其他字段（不支持复合主键，必须使用单字段唯一性）。
  • “认证过期” → 在数据源中轮换凭证，然后重新测试连接。
  • 流卡在处理中状态 → 查看连接器页面了解源端速率限制；验证源端调度设置。
<!-- SF_DOC_SYNC_END:connect-data-help-side --> <!-- SF_DOC_SYNC_START:developer-integration-catalog -->

Integration and connector catalog gate

集成与连接器目录关卡

Auto-synced from sf-docs captures of official Salesforce Developer documentation.
Sources:
Source fingerprint: 
189417577b6f42e77f931e75
Implementation notes:
  • Classify each connector by supported direction and mode: ingestion, query federation, file federation, data share, unstructured ingestion, activation, or bidirectional use.
  • For Databricks, select the exact mode before setup; batch ingestion, query federation, file federation, and data sharing have different network, compute, catalog, and storage proof paths.
  • For file federation, verify both the catalog endpoint and underlying object storage path, supported table format, source table eligibility, and required grants.
  • For Ingestion API, treat schema agreement, connector setup, External Client App auth, data-stream deployment, object-endpoint delivery, and DMO mapping as separate gates.
  • Connector availability, authentication, limitations, and supported objects change frequently; route current claims back through the exact connector page and Help limits.
<!-- SF_DOC_SYNC_END:developer-integration-catalog -->
自动同步自sf-docs抓取的Salesforce官方开发者文档。
来源文档：
源指纹： 
189417577b6f42e77f931e75
实施说明：
  • 按支持的方向和模式对每个连接器进行分类：ingestion、查询联邦、文件联邦、数据共享、非结构化数据摄入、激活或双向使用。
  • 对于Databricks，在设置前选择确切模式；批量ingestion、查询联邦、文件联邦和数据共享有不同的网络、计算、目录和存储验证路径。
  • 对于文件联邦，验证目录端点和底层对象存储路径、支持的表格式、源表资格以及所需权限。
  • 对于Ingestion API，将schema一致性、连接器设置、外部客户端应用认证、数据流部署、对象端点交付和DMO映射视为独立的验证关卡。
  • 连接器的可用性、认证方式、限制和支持对象会频繁变化；当前相关声明需通过确切的连接器页面和帮助文档中的限制进行确认。
<!-- SF_DOC_SYNC_END:developer-integration-catalog -->