Managing Event Specifications via the API
With the Event Specifications API, you can efficiently manage event specifications programmatically. Whether you want to retrieve, create, edit, publish, deprecate, or delete event specifications, the API provides the necessary endpoints and functionalities.
In the current API, we refer to all data and entry points for event specifications as "tracking scenarios." We're working on a new version, and in a future update, we'll use the name "event specifications" for them.
Getting Startedโ
For comprehensive details regarding each request, including query parameters and response formats, please consult the relevant API documentation.
Authorizing in the API Documentationโ
To post sample requests in the documentation, you must click the Authorize
button at the top of the document and authenticate with your token. The token field value in each individual request is overwritten by this authorization.
For each request, you are required to include your company's organizationId
, which is a UUID retrievable from the URL immediately following .com
when visiting the console:
Obtaining an API keyโ
Create a new api key for your organization in console.
Use the generated api key to obtain an authorization token.
curl \
--header 'X-API-Key: $API_KEY' \
https://console.snowplowanalytics.com/api/msc/v1/organizations/$ORGANIZATION_ID/credentials/v2/token
This command will return an access token wrapped in json.
{"accessToken":"<access token value>"}
You may then use this access token value to supply authorization headers for subsequent api requests.
curl \
--header 'authorization: Bearer $ACCESS_TOKEN_VALUE'
Response Formatโ
The tracking scenarios API follows a specific response format for successful cases (2xx
) and for scenarios where critical and non-critical errors may occur, such as (422
) Unprocessable Entity.
{
"data": [
// Tracking scenarios
],
"includes": [
// Additional tracking scenario info
],
"errors": [
// Warnings or errors
]
}
data
: Contains the scenario or scenarios, depending on the requestincludes
: Provides additional information, such as the history of tracking scenario changeserrors
: Holds a list of errors, which could be of typeError
orWarning
. If the array field contains at least one error of typeError
, the request will also return a4xx
status code, indicating that it cannot perform the store operation. Any other severity different fromError
will result in a2xx
status code.
Compatibility Checksโ
Certain endpoints conduct a validation to verify the compatibility of a specific tracking scenario event schema, event.schema
, with the source data structure version referenced by event.source
. When both event.schema
and event.source
are defined in the tracking scenario, the compatibility checks will be executed.
...
"event": {
"source": "iglu:com.example/ui_actions/jsonschema/1-0-0",
"schema": {
"$schema": "http://json-schema.org/draft-04/schema#",
"description": "Event to capture search submit",
"type": "object",
"required": [
"label",
"action"
],
"properties": {
"action": {
"type": "string",
"enum": [
"click"
]
},
"label": {
"type": "string",
"enum": [
"Search"
]
}
},
"additionalProperties": false
}
}
...
However, the compatibility check will not only be performed against the version specified by the source data structure (event.source
field, e.g., 1-0-0
), which we will refer to as the current version. It will also be conducted against the latest version available in Iglu, referred to as the latest version.
This approach is because it's common for a new tracking scenario to utilize the latest version of the source data structure. However, as this data structure may evolve over time and become incompatible with the event.schema
defined in the tracking scenario, we provide a method to detect these compatibility issues. Consequently, customers can update the tracking scenario to ensure compatibility.
In cases where a scenario is incompatible or we cannot determine it, some errors will be provided in the errors
field of the response. These errors alerting of compatibility issues between the tracking scenario and the source data structure will take a similar shape to the one below:
...
"errors": [
{
"type":"Warning",
"code":"SchemaIncompatible",
"title":"Tracking scenario with id: 59b5e250-91c4-45af-a63d-5f8cd39f4b67, event schema is INCOMPATIBLE with schema with name: test_event, vendor: com.snplow.msc.aws, version: 1-0-13",
"source":"event.schema"
}
]
...
Compatibility checks can result in three possible values: Compatible, SchemaIncompatible, or SchemaUndecidable.
- If Compatible, the tracking scenario is compatible, and no errors will be appended to the
errors
response field - If SchemaIncompatible, the tracking scenario is incompatible against some version. If the check for the current version is incompatible, the
type
will beError
. For incompatibility with the latest version, thetype
will beWarning
. If the requested operation involves persisting the scenario (create/update), an error of typeError
will be appended to the response, the status code will be 422 Unprocessable Entity, and the store operation will not persist. When fetching a scenario, the checks will run for both too, current and latest versions, and if incompatible, the error type will always beWarning
, returning status code 200 Ok. - If SchemaUndecidable, it is indeterminable whether the scenario is compatible with a specific version due to the use of some advanced JSON-Schema features and the high computational cost of checking compatibility. The
type
will always beWarning
, and the user is responsible for ensuring that the tracking scenario is compatible with the source data structure. A warning will be attached to theerrors
response field.
The algorithm used to perform the compatibility check is based on the Finding Data Compatibility Bugs with JSON Subschema Checking paper, authored by Andrew Habib, Avraham Shinnar, and Michael Pradel.
Retrieve a List of Tracking Scenariosโ
Use this request to retrieve a list of scenarios within an organization, which will be wrapped into the data
field of the response.
GET /api/msc/v1/organizations/{organizationId}/tracking-scenarios/v2
The organizationId
parameter is required.
Query Parameters and Filtersโ
You can filter the results based on the following query parameters:
dataProductId
: Filters the tracking scenarios associated with a particular data productsourceId
: Filters the tracking scenarios associated with a particular data structure, inferred from theevent.source
fieldsourceVersion
: Filters the tracking scenarios associated with a specific data structure version when used withdataStructureId
.withLatestHistory
: Whentrue
, it will return a list of tracking scenarios, with the latest change per scenario attached to theincludes
array field. The relation between tracking scenarios indata
and history inincludes
can be determined byid = scenarioId
.status
: Filters the tracking scenarios that match the specified status
If no query parameters are provided, it will return all the tracking scenarios for an organization.
Retrieve a Specific Tracking Scenarioโ
Use this request to retrieve a specific tracking scenario within an organization. The retrieved scenario will be wrapped into the data
field of the response.
GET /api/msc/v1/organizations/{organizationId}/tracking-scenarios/v2/{scenarioId}
This endpoint will trigger compatibility checking if event.source
and event.schema
are defined.
Query parameters organizationId
and scenarioId
are required:
withHistory
: Whentrue
, returns a list with the history for the tracking scenario in theincludes
array field of the response, related to the tracking scenario by its idstatus
: Filters the tracking scenarios that match the specified status.
Creating a Tracking Scenarioโ
Use this request to create a tracking scenario within an organization.
POST /api/msc/v1/organizations/{organizationId}/tracking-scenarios/v2
Query parameter organizationId
is required.
Request body example:
{
"scenario": {
"name": "Search",
"description": "Tracking the use of the search box",
"event": {
"source": "iglu:com.example/ui_actions/jsonschema/1-0-0"
}
},
"message": "update"
}
The creation form has two fields at the top level, as shown in the example above:
message
: An optional field to provide a messagescenario
: The definition of the tracking scenario, which should comply with the validations.
By default, the tracking scenario will be created with scenario.status
set to draft
and scenario.version
set to 0
if not provided. These values can be changed and managed after creation. Here is an example response:
{
"data": [
{
"id": "5a203ef8-939b-4fd1-914e-f12a3dd1a869",
"version": 0,
"status": "draft",
"name": "Search",
"description": "Tracking the use of the search box",
"event": {
"source": "iglu:com.example/ui_actions/jsonschema/1-0-0"
}
}
],
"includes": [
{
"author": "39b81015-1bd5-4b37-96c7-3296cabaa36f",
"message": "initial draft",
"date": "2023-04-26T14:41:48.708191Z",
"scenarioId": "5a203ef8-939b-4fd1-914e-f12a3dd1a869",
"version": 0,
"status": "draft",
"type": "History"
}
],
"errors": []
}
Validationsโ
scenario.event.source
: If provided it should match a valid and existing Iglu URIscenario.name
: It validates that thescenario.name
of a tracking scenario is unique within the data structure context, inferred from the source data structurescenario.event.source
if providedscenario.version
: If provided should be equal or greater than zeroscenario.status
: If provided should match one ofdraft
,published
ordeprecated
scenario.entities
: If provided it will validate that the entities,scenario.entities.tracked
andscenario.entities.enriched
, are not duplicated and that they existscneario.dataProductId
: If provided it will validate that the data products exists. (Coming soon)
This endpoint will trigger compatibility checking if event.source
and event.schema
are defined.
Editing a Tracking Scenarioโ
Use this request to edit a tracking scenario within an organization. The format of the request and response is the same as during creation.
The organizationId
and scenarioId
parameters are required.
PUT /api/msc/v1/organizations/{organizationId}/tracking-scenarios/v2/{scenarioId}
Publishing a Tracking Scenarioโ
When editing a scenario, it can be published by setting the status
to published
. Currently, this will indicate to the tracking scenario consumers (for instance, front-end developers) that the tracking design is ready to be implemented or consumed.
By default, when a tracking scenario is created and no value is provided for scenario.status
, it will be set to draft
. With this, we suggest a tracking scenario lifecycle that we recommend following, but we allow a certain degree of flexibility to accommodate unique customer use cases. Here is the suggested lifecycle:
In addition to this lifecycle, and in conjunction with versioning, we enforce that when a tracking scenario is published, the versions between two published versions are discarded. For example:
Publish new version, before squash:
After discarding intermediate versions:
Deprecating a Tracking Scenarioโ
When editing a scenario, it can be deprecated by setting the status
to deprecated
. This is a way of informing tracking scenario consumers (for instance, developers) not to rely on the tracking anymore.
Validationsโ
scenario.event.source
: If provided, it should match a valid and existing Iglu URIscenario.name
: It validates that thescenario.name
of a tracking scenario is unique within the data structure context, inferred from the source data structurescenario.event.source
if providedscenario.version
: If provided, it should be equal to or greater than zero, should not exist, and be greater than the last published versionscenario.status
: If provided, it should match one ofdraft
,published
, ordeprecated
scenario.entities
: If provided, it will validate that the entities,scenario.entities.tracked
andscenario.entities.enriched
, are not duplicated and that they existscenario.dataProductId
: If provided, it will validate that the data product exists
This endpoint will trigger compatibility checking if event.source
and event.schema
are defined.
Deleting a Tracking Scenarioโ
Use this request to delete a tracking scenario within an organization.
DELETE /api/msc/v1/organizations/{organizationId}/tracking-scenarios/v2/{scenarioId}
Please note that this action is irreversible and will permanently delete the tracking scenario.