Manage an Endpoint

An endpoint in Data Service (beta) is a web API that you can customize to execute SQL statements. You can specify parameters for the SQL statements, such as the value used in the WHERE clause. When a client calls an endpoint and provides values for the parameters in a request URL, the endpoint executes the SQL statement with the provided parameters and returns the results as part of the HTTP response.

This document describes how to manage your endpoints in a Data App in the TiDB Cloud console.

Before you begin

  • Before you create an endpoint, make sure the following:

    • You have created a cluster and a Data App. For more information, see Create a Data App.
    • The databases, tables, and columns that the endpoint will operate on already exist in the target cluster.
  • Before you call an endpoint, make sure that you have created an API key in the Data App. For more information, see Create an API key.

Create an endpoint

In Data Service, you can automatically generate endpoints, manually create endpoints, or add predefined system endpoints.

Generate an endpoint automatically

In TiDB Cloud Data Service, you can generate one or multiple endpoints automatically in one go as follows:

  1. Navigate to the Data Service page of your project.

  2. In the left pane, locate your target Data App, click + to the right of the App name, and then click Autogenerate Endpoint. The dialog for endpoint generation is displayed.

  3. In the dialog, do the following:

    1. Select the target cluster, database, and table for the endpoint to be generated.

    2. Select at least one HTTP operation (such as GET (Retrieve), POST (Create), and PUT (Update)) for the endpoint to be generated.

      For each operation you select, TiDB Cloud Data Service will generate a corresponding endpoint. If you select a batch operation (such as POST (Batch Create)), the generated endpoint lets you operate on multiple rows in a single request.

      If the table you selected contains vector data types, you can enable the Vector Search Operations option and select a vector distance function to generate a vector search endpoint that automatically calculates vector distances based on your selected distance function. The supported vector distance functions include the following:

      • VEC_L2_DISTANCE (default): calculates the L2 distance (Euclidean distance) between two vectors.
      • VEC_COSINE_DISTANCE: calculates the cosine distance between two vectors.
      • VEC_NEGATIVE_INNER_PRODUCT: calculates the distance by using the negative of the inner product between two vectors.
      • VEC_L1_DISTANCE: calculates the L1 distance (Manhattan distance) between two vectors.
    3. (Optional) Configure a timeout and tag for the operations. All the generated endpoints will automatically inherit the configured properties, which can be modified later as needed.

    4. (Optional) The Auto-Deploy Endpoint option (disabled by default) controls whether to enable the direct deployment of the generated endpoints. When it is enabled, the draft review process is skipped, and the generated endpoints are deployed immediately without further manual review or approval.

  4. Click Generate.

    The generated endpoint is displayed at the top of the endpoint list.

  5. Check the generated endpoint name, SQL statements, properties, and parameters of the new endpoint.

    • Endpoint name: the generated endpoint name is in the /<name of the selected table> format, and the request method (such as GET, POST, and PUT) is displayed before the endpoint name. For example, if the selected table name is sample_table and the selected operation is POST (Create), the generated endpoint is displayed as POST /sample_table.

      • If a batch operation is selected, TiDB Cloud Data Service appends /bulk to the name of the generated endpoint. For example, if the selected table name is /sample_table and the selected operation is POST (Batch Create), the generated endpoint is displayed as POST /sample_table/bulk.
      • If POST (Vector Similarity Search) is selected, TiDB Cloud Data Service appends /vector_search to the name of the generated endpoint. For example, if the selected table name is /sample_table and the selected operation is POST (Vector Similarity Search), the generated endpoint is displayed as POST /sample_table/vector_search.
      • If there has been already an endpoint with the same request method and endpoint name, TiDB Cloud Data Service appends _dump_<random letters> to the name of the generated endpoint. For example, /sample_table_dump_EUKRfl.
    • SQL statements: TiDB Cloud Data Service automatically writes SQL statements for the generated endpoints according to the table column specifications and the selected endpoint operations. You can click the endpoint name to view its SQL statements in the middle section of the page.

    • Endpoint properties: TiDB Cloud Data Service automatically configures the endpoint path, request method, timeout, and tag according to your selection. You can find the properties in the right pane of the page.

    • Endpoint parameters: TiDB Cloud Data Service automatically configures parameters for the generated endpoints. You can find the parameters in the right pane of the page.

  6. If you want to modify the details of the generated endpoint, such as its name, SQL statements, properties, or parameters, refer to the instructions provided in Develop an endpoint.

Create an endpoint manually

To create an endpoint manually, perform the following steps:

  1. Navigate to the Data Service page of your project.
  2. In the left pane, locate your target Data App, click + to the right of the App name, and then click Create Endpoint.
  3. Update the default name if necessary. The newly created endpoint is added to the top of the endpoint list.
  4. Configure the new endpoint according to the instructions in Develop an endpoint.

Add a predefined system endpoint

Data Service provides an endpoint library with predefined system endpoints that you can directly add to your Data App, reducing the effort in your endpoint development. Currently, the library only includes the /system/query endpoint, which enables you to execute any SQL statement by simply passing the statement in the predefined sql parameter.

To add a predefined system endpoint to your Data App, perform the following steps:

  1. Navigate to the Data Service page of your project.

  2. In the left pane, locate your target Data App, click + to the right of the App name, and then click Manage Endpoint Library.

    A dialog for endpoint library management is displayed. Currently, only Execute Query (that is, the /system/query endpoint) is provided in the dialog.

  3. To add the /system/query endpoint to your Data App, toggle the Execute Query switch to Added.

  4. Click Save.

    The added system-provided endpoint is displayed at the top of the endpoint list.

  5. Check the endpoint name, SQL statements, properties, and parameters of the new endpoint.

    • Endpoint name: the endpoint name and path is /system/query, and the request method POST.
    • SQL statements: the /system/query endpoint does not come with any SQL statement. You can find the SQL editor in the middle section of the page and write your desired SQL statements in the SQL editor. Note that the SQL statements written in the SQL editor for the /system/query endpoint will be saved in the SQL editor so you can further develop and test them next time but they will not be saved in the endpoint configuration.
    • Endpoint properties: in the right pane of the page, you can find the endpoint properties on the Properties tab. Unlike other custom endpoints, only the timeout and max rows properties can be customized for system endpoints.
    • Endpoint parameters: in the right pane of the page, you can find the endpoint parameters on the Params tab. The parameters of the /system/query endpoint are configured automatically and cannot be modified.

Develop an endpoint

For each endpoint, you can write SQL statements to execute on a TiDB cluster, define parameters for the SQL statements, or manage the name and version.

Configure properties

On the right pane of the endpoint details page, you can click the Properties tab to view and configure properties of the endpoint.

Basic properties

  • Path: the path that users use to access the endpoint.

    • The length of the path must be less than 64 characters.

    • The combination of the request method and the path must be unique within a Data App.

    • Only letters, numbers, underscores (_), slashes (/), and parameters enclosed in curly braces (such as {var}) are allowed in a path. Each path must start with a slash (/) and end with a letter, number, or underscore (_). For example, /my_endpoint/get_id.

    • For parameters enclosed in { }, only letters, numbers, and underscores (_) are allowed. Each parameter enclosed in { } must start with a letter or underscore (_).

  • Endpoint URL: (read-only) the default URL is automatically generated based on the region where the corresponding cluster is located, the service URL of the Data App, and the path of the endpoint. For example, if the path of the endpoint is /my_endpoint/get_id, the endpoint URL is https://<region>.data.tidbcloud.com/api/v1beta/app/<App ID>/endpoint/my_endpoint/get_id. To configure a custom domain for the Data App, see Custom Domain in Data Service.

  • Request Method: the HTTP method of the endpoint. The following methods are supported:

    • GET: use this method to query or retrieve data, such as a SELECT statement.
    • POST: use this method to insert or create data, such as an INSERT statement.
    • PUT: use this method to update or modify data, such as an UPDATE statement.
    • DELETE: use this method to delete data, such as a DELETE statement.
  • Description (Optional): the description of the endpoint.

Advanced properties

  • Timeout(ms): the timeout for the endpoint, in milliseconds.

  • Max Rows: the maximum number of rows that the endpoint can operate or return.

  • Tag: the tag used for identifying a group of endpoints.

  • Pagination: this property is available only when the request method is GET and the last SQL statement of the endpoint is a SELECT operation. When Pagination is enabled, you can paginate the results by specifying page and page_size as query parameters when calling the endpoint, such as https://<region>.data.tidbcloud.com/api/v1beta/app/<App ID>/endpoint/my_endpoint/get_id?page=<Page Number>&page_size=<Page Size>. For more information, see Call an endpoint.

  • Cache Response: this property is available only when the request method is GET. When Cache Response is enabled, TiDB Cloud Data Service can cache the response returned by your GET requests within a specified time-to-live (TTL) period.

  • Time-to-live(s): this property is available only when Cache Response is enabled. You can use it to specify the time-to-live (TTL) period in seconds for cached response. During the TTL period, if you make the same GET requests again, Data Service returns the cached response directly instead of fetching data from the target database again, which improves your query performance.

  • Batch Operation: this property is visible only when the request method is POST or PUT. When Batch Operation is enabled, you can operate on multiple rows in a single request. For example, you can insert multiple rows of data in a single POST request by putting an array of data objects to the items field of an object in the --data-raw option of your curl command when calling the endpoint.

Write SQL statements

On the SQL editor of the endpoint details page, you can write and run the SQL statements for an endpoint. You can also simply type -- followed by your instructions to let AI generate SQL statements automatically.

  1. Select a cluster.

    On the upper part of the SQL editor, select a cluster on which you want to execute SQL statements from the drop-down list. Then, you can view all databases of this cluster in the Schema tab on the right pane.

  2. Depending on your endpoint type, do one of the following to select a database:

    • Predefined system endpoints: on the upper part of the SQL editor, select the target database from the drop-down list.
    • Other endpoints: write a SQL statement to specify the target database in the SQL editor. For example, USE database_name;.
  3. Write SQL statements.

    In the SQL editor, you can write statements such as table join queries, complex queries, and aggregate functions. You can also simply type -- followed by your instructions to let AI generate SQL statements automatically.

    To define a parameter, you can insert it as a variable placeholder like ${ID} in the SQL statement. For example, SELECT * FROM table_name WHERE id = ${ID}. Then, you can click the Params tab on the right pane to change the parameter definition and test values. For more information, see Parameters.

    When defining an array parameter, the parameter is automatically converted to multiple comma-separated values in the SQL statement. To make sure that the SQL statement is valid, you need to add parentheses (()) around the parameter in some SQL statements (such as IN). For example, if you define an array parameter ID with test value 1,2,3, use SELECT * FROM table_name WHERE id IN (${ID}) to query the data.

  4. Run SQL statements.

    If you have inserted parameters in the SQL statements, make sure that you have set test values or default values for the parameters in the Params tab on the right pane. Otherwise, an error is returned.

    • macOS
    • Windows/Linux

    For macOS:

    • If you have only one statement in the editor, to run it, press ⌘ + Enter or click Run.

    • If you have multiple statements in the editor, to run one or several of them sequentially, place your cursor on your target statement or select the lines of the target statements with your cursor, and then press ⌘ + Enter or click Run.

    • To run all statements in the editor sequentially, press ⇧ + ⌘ + Enter, or select the lines of all statements with your cursor and click Run.

    For Windows or Linux:

    • If you have only one statement in the editor, to run it, press Ctrl + Enter or click Run.

    • If you have multiple statements in the editor, to run one or several of them sequentially, place your cursor on your target statement or select the lines of the target statements with your cursor, and then press Ctrl + Enter or click Run.

    • To run all statements in the editor sequentially, press Shift + Ctrl + Enter, or select the lines of all statements with your cursor and click Run.

    After running the statements, you can see the query results immediately in the Result tab at the bottom of the page.

Configure parameters

On the right pane of the endpoint details page, you can click the Params tab to view and manage the parameters used in the endpoint.

In the Definition section, you can view and manage the following properties for a parameter:

  • The parameter name: the name can only include letters, digits, and underscores (_) and must start with a letter or an underscore (_). DO NOT use page and page_size as parameter names, which are reserved for pagination of request results.

  • Required: specifies whether the parameter is required in the request. For path parameters, the configuration is required and cannot be modified. For other parameters, the default configuration is not required.

  • Type: specifies the data type of the parameter. For path parameters, only STRING and INTEGER are supported. For other parameters, STRING, NUMBER, INTEGER, BOOLEAN, and ARRAY are supported.

    When using a STRING type parameter, you do not need to add quotation marks (' or "). For example, foo is valid for the STRING type and is processed as "foo", whereas "foo" is processed as "\"foo\"".

  • Enum Value: (optional) specifies the valid values for the parameter and is available only when the parameter type is STRING, INTEGER, or NUMBER.

    • If you leave this field empty, the parameter can be any value of the specified type.
    • To specify multiple valid values, you can separate them with a comma (,). For example, if you set the parameter type to STRING and specify this field as foo, bar, the parameter value can only be foo or bar.
  • ItemType: specifies the item type of an ARRAY type parameter.

  • Default Value: specifies the default value of the parameter.

    • For ARRAY type, you need to separate multiple values with a comma (,).
    • Make sure that the value can be converted to the type of parameter. Otherwise, the endpoint returns an error.
    • If you do not set a test value for a parameter, the default value is used when testing the endpoint.
  • Location: indicates the location of the parameter. This property cannot be modified.

    • For path parameters, this property is Path.
    • For other parameters, if the request method is GET or DELETE, this property is Query. If the request method is POST or PUT, this property is Body.

In the Test Values section, you can view and set test parameters. These values are used as the parameter values when you test the endpoint. Make sure that the value can be converted to the type of parameter. Otherwise, the endpoint returns an error.

Rename

To rename an endpoint, perform the following steps:

  1. Navigate to the Data Service page of your project.
  2. In the left pane, click the name of your target Data App to view its endpoints.
  3. Locate the endpoint you want to rename, click ... > Rename., and enter a new name for the endpoint.

Test an endpoint

To test an endpoint, perform the following steps:

  1. Navigate to the Data Service page of your project.

  2. In the left pane, click the name of your target Data App to view its endpoints.

  3. Click the name of the endpoint you want to test to view its details.

  4. (Optional) If the endpoint contains parameters, you need to set test values before testing.

    1. On the right pane of the endpoint details page, click the Params tab.

    2. Expand the Test Values section and set test values for the parameters.

      If you do not set a test value for a parameter, the default value is used.

  5. Click Test in the upper-right corner.

After testing the endpoint, you can see the response as JSON at the bottom of the page. For more information about the JSON response, refer to Response of an endpoint.

Deploy an endpoint

To deploy an endpoint, perform the following steps:

  1. Navigate to the Data Service page of your project.

  2. In the left pane, click the name of your target Data App to view its endpoints.

  3. Locate the endpoint you want to deploy, click the endpoint name to view its details, and then click Deploy in the upper-right corner.

  4. If Review Draft is enabled for your Data App, a dialog is displayed for you to review the changes you made. You can choose whether to discard the changes based on the review.

  5. Click Deploy to confirm the deployment. You will get the Endpoint has been deployed prompt if the endpoint is successfully deployed.

    On the right pane of the endpoint details page, you can click the Deployments tab to view the deployed history.

Call an endpoint

To call an endpoint, you can send an HTTPS request to either an undeployed draft version or a deployed online version of the endpoint.

Prerequisites

Before calling an endpoint, you need to create an API key. For more information, refer to Create an API key.

Request

TiDB Cloud Data Service generates code examples to help you call an endpoint. To get the code example, perform the following steps:

  1. Navigate to the Data Service page of your project.

  2. In the left pane, click the name of your target Data App to view its endpoints.

  3. Locate the endpoint you want to call and click ... > Code Example. The Code Example dialog box is displayed.

  4. In the dialog box, select the environment and authentication method that you want to use to call the endpoint, and then copy the code example.

    • Environment: choose Test Environment or Online Environment depending on your need. Online Environment is available only after you deploy the endpoint.

    • Authentication method: choose Basic Authentication or Digest Authentication.

      • Basic Authentication transmits your API key as based64 encoded text.
      • Digest Authentication transmits your API key in an encrypted form, which is more secure.

      Compared with Basic Authentication, the curl code of Digest Authentication includes an additional --digest option.

    Here is an example of a curl code snippet for a POST request that enables Batch Operation and uses Digest Authentication:

    • Test Environment
    • Online Environment

    To call a draft version of the endpoint, you need to add the endpoint-type: draft header:

    curl --digest --user '<Public Key>:<Private Key>' \ --request POST 'https://<region>.data.tidbcloud.com/api/v1beta/app/<App ID>/endpoint/<Endpoint Path>' \ --header 'content-type: application/json'\ --header 'endpoint-type: draft' --data-raw '{ "items": [ { "age": "${age}", "career": "${career}" } ] }'

    You must deploy your endpoint first before checking the code example in the online environment.

    To call the current online version of the endpoint, use the following command:

    curl --digest --user '<Public Key>:<Private Key>' \ --request POST 'https://<region>.data.tidbcloud.com/api/v1beta/app/<App ID>/endpoint/<Endpoint Path>' \ --header 'content-type: application/json'\ --data-raw '{ "items": [ { "age": "${age}", "career": "${career}" } ] }'
  5. Paste the code example in your application, edit the example according to your need, and then run it.

    • You need to replace the <Public Key> and <Private Key> placeholders with your API key. For more information, refer to Manage an API key.

    • If the request method of your endpoint is GET and Pagination is enabled for the endpoint, you can paginate the results by updating the values of page=<Page Number> and page_size=<Page Size> with your desired values. For example, to get the second page with 10 items per page, use page=2 and page_size=10.

    • If the request method of your endpoint is POST or PUT, fill in the --data-raw option according to the rows of data that you want to operate on.

      • For endpoints with Batch Operation enabled, the --data-raw option accepts an object with an items field containing an array of data objects so you can operate on multiple rows of data using one endpoint.
      • For endpoints with Batch Operation not enabled, the --data-raw option only accepts one data object.
    • If the endpoint contains parameters, specify the parameter values when calling the endpoint.

Response

After calling an endpoint, you can see the response in JSON format. For more information, see Response and Status Codes of Data Service.

Undeploy an endpoint

To undeploy an endpoint, perform the following steps:

  1. Navigate to the Data Service page of your project.
  2. In the left pane, click the name of your target Data App to view its endpoints.
  3. Locate the endpoint you want to undeploy, click ... > Undeploy.
  4. Click Undeploy to confirm the undeployment.

Delete an endpoint

To delete an endpoint, perform the following steps:

  1. Navigate to the Data Service page of your project.
  2. In the left pane, click the name of your target Data App to view its endpoints.
  3. Click the name of the endpoint you want to delete, and then click ... > Delete in the upper-right corner.
  4. Click Delete to confirm the deletion.

Was this page helpful?