What are customer-configurable event destinations?

An event destination is any system a product event can be delivered to: a webhook endpoint, a storage bucket, a warehouse, a queue, a database, or an analytics tool. "Customer-configurable" means the customer, not the vendor's engineering team, decides which destinations exist, supplies the credentials, and sets the rules for what gets delivered where.

In this pattern, the vendor ships destination types as capabilities and exposes configuration to the customer. One customer might send order events to a webhook and a warehouse; another might route only high-value events to a queue. The same delivery code path serves all of them because the differences live in configuration, not in bespoke integration code.

When integrations are hardcoded, every new customer request becomes an engineering ticket. A customer wants their events in a different warehouse, on their own credentials, filtered to a subset of event types, and each variation means code, review, deploy, and ongoing maintenance. The backlog of one-off integrations grows with the customer base.

Hardcoding also concentrates operational load on the vendor. The vendor holds the credentials, fields every "why didn't my event arrive?" question, and owns failure recovery for integrations it built per account. That does not scale, and it makes the vendor the bottleneck for setup, rotation, and triage.

Making destinations genuinely configurable takes more than a dropdown. Each destination type needs its own authentication model and credential handling, with secrets stored encrypted and masked on display. Customers need routing, choosing which event types go to which destination, and basic field filtering so they receive only what they want.

It also requires observability the customer can use directly: a delivery log, the payload that was sent, the response that came back, the reason a delivery failed, and a way to replay failures once the destination is healthy. Without that, "configurable" stops at setup and leaves the customer blind the moment something breaks.

Pushrail makes event destinations customer-configurable across its full catalog, webhooks, object storage, data warehouses, message queues, streams, databases, and analytics tools. Customers configure destinations from a dashboard or an embedded portal, supply their own credentials, define routing rules, and apply basic filtering and lightweight transforms.

Reliability and observability come with it: deliveries are retried with backoff, classified as transient or permanent, dead-letter-queued when exhausted, logged per attempt, and replayable. Credentials are encrypted and scoped to the customer that owns them, and tenant isolation keeps one customer's configuration and deliveries separate from another's.