The use case for Firehose differs from REST and Sync APIs due to its key features:
Provides near real-time integration with high volumes of data
Captures granular details around each event and data change
Delivers a complete, sequential stream of events, rather than just the most recent
The Firehose API can be used to power automations that support complex and unique sales processes within Base and other business systems, including marketing automation platforms, messaging applications, quoting/contracting solutions and more. These workflows save reps significant time and help close more deals faster by effectively facilitating the sales process and highlighting priorities.
For example, Firehose can be used to change a contact’s deal stage in Base or create a task for the rep to follow up again at a specific time based on the outcome of a call. Similarly, if a deal is moved to a particular stage in Base, Firehose can trigger events in other systems such as sending an email or text message, dynamically generating a proposal or creating a support ticket.
Power Business Intelligence
Firehose can also pass a continuous stream of events directly to business intelligence and sales analytics platforms. Event-based data aids in identifying the types of activities that convert the most successfully, the ideal number of touches needed to close a deal, and other actionable insights. It can also be used to fuel real-time sales team performance dashboards.
Finally, Firehose can be used to generate alerts within Base and external systems based on granular information in Base. For instance, Firehose can power a hot lead notification in Hipchat or Slack, or a reminder via push notification to schedule a meeting or make a call.
Firehose API is designed around the concept of resources, such as a lead, contact or deal.
Every resource has its own Firehose endpoint that is used to receive events for particular resource.
Each response returns a limited collection of events. The client needs to iteratively call the endpoint in order to drain the event queue.
Event Stream Endpoint
This endpoint provides a stream of changes to Base data. It retains full event history from the last 72 hours, which you can use to back fill your application if it has been down for some time.
All requests to the Firehose API must be authenticated through the standard Authorization header using the
Bearer authentication scheme to transmit the access token. See the OAuth 2.0 protocol for more details.
Firehose API user
Important: You can use Firehose API only in the context of the root user.
Root user has complete data access on your Base account.
To utilize Firehose API you will need to activate Tree Based Permissions on your account and then generate an OAuth Token from your root user’s Base account.
Firehose API accepts query parameters described in the table below.
Your client position in Firehose stream.
Limits maximum number of events in single response.
The position parameter is required when calling Firehose API. It indicates current client position in the stream that will be used to start consuming events from.
When calling Firehose API for the first time you may start from the end of the queue using the top position (newest events).
Every response will return the next position in the stream, which you will use to make the subsequent call.
Notice that this parameter only enforces the upper limit of events returned in single response.
It is not guaranteed to return as many events as the value of the limit in a single response, even if additional data is available.
Firehose API responses contain collections of events for a particular resource.
Each event provides a current snapshot of data along with information regarding what has changed in this singular event and which user enacted the change.
A collection of objects the payload carries.
Payload’s metadata information.
A collection of error objects.
In a successful response, the meta object holds 3 fields:
Link to retrieve next data chunk.
Tells whether you reached the current end of the stream.
Next position to start consuming events from.
This indicator alerts you as to whether or not there is more data available to consume.
Notice that new data may appear any time and you should call API periodically even if top is equal to true.
Represents your client position in the stream of changes.
It is the client’s responsibility to store the position value. Your workers may fail, so you should retain your position value
in non-volatile memory (e.g. a database) in order to begin from the previous position after resumption.
Otherwise, you will have to re-initialize your client’s position again.
See Consumption Strategies section for more details.
On an error response, the metadata object will hold error-related information such as the http_status and logref attributes.
The items field is always an JSON array. An item represents a single event for the resource. Detailed format of objects depends on which resource type you are requesting.
Each event has standard format that includes two separate sections: data and meta.
Contains current snapshot of the data. See documentation for particular resource to see what data is returned.
Contains various contextual attributes:
Unique id generated for every event (e.g. useful for tracing)
Resource event type, one of: created, updated, deleted.
Monotonic sequence aligned with order of resource events.
Contains previous values for all the properties that have been modified within particular event. (available only when event_type equals updated).
Firehose will attach increasing number (may be not continuous) for each new resource event for particular resource id.
It may be used to check which one from two events for same resource id is newer (e.g. to prevent old data processing).
Events are typically generated by user interactions. In some cases, events will also be generated by the system itself.
interaction - user action.
system - admin action (e.g. initial bootstrap) or a backward compatible format change (e.g. new attribute added).
embedded_object_event - Firehose will join some resources together (e.g. custom field value in a deal resource references the definition of custom field). In case of delete/update of embedded resource, e.g. if a custom field definition has been deleted/updated, the user may be interested to see the new event for the deal that reflects the change, i.e. a deal with a newly updated name of the custom field. Firehose guarantees that after the embedded object event all subsequent events for the primary resource will contain latest state of the embedded object.
Not all event_cause’s are possible in all event_type’s:
interaction, system, embedded_object_event
Important: When calling API using top position, items array is always empty and top is set to true. This is because top request is just used to get current position in the stream of events. This position needs to be used in a subsequent call to get most recent data since the top request was issued.
Pricing And Rate Limiting
The Firehose API is a premium Base API that is part of the Snap Advanced add-on. Checkout pricing details.