Server-Sent Events (SSE)
As soon as you deploy an event source, Pipedream creates a private SSE stream linked to that source. Each event emitted by the source is published to this SSE stream, so your app can subscribe to the stream to process new events.
What is SSE?
Unlike WebSockets, SSE enables one-way communication from server to clients (WebSockets enable bidirectional communication between server and client, allowing you to pass messages back and forth). If you only need a client to subscribe to events from a server, and don't require bidirectional communication, SSE is simple way to make that happen.
Connecting to the SSE stream directly
To process events from your source's SSE stream, you'll need to:
- Get the SSE URL for your source's events using
pd list streams. You'll see your stream's URL under the SSE header.
- Connect to the SSE stream, passing your Pipedream API key in the
AuthorizationHTTP header using Bearer Auth.
See this repo for an example Node.js app that processes events from a Pipedream SSE stream.
Most programming languages provide SSE clients that facilitate interaction with SSE streams. For example, the Node.js example repo uses the
eventsource npm package, which implements the
Subscribe to new events using the Pipedream CLI
pd events command connects to the SSE stream for a source and prints new events as they're emitted:
pd events -f <source-id-or-name>
This is the easiest way to interact with streams, especially if you only need to manually inspect events, or just want to save them to a file on disk.
Every 15 seconds, we'll send a message with the
:sse-handshake comment to keep open SSE connections alive. These comments should be automatically ignored when you're listening for messages using the
We send these because the SSE spec notes that
Legacy proxy servers are known to, in certain cases, drop HTTP connections after a short timeout. To protect against such proxy servers, authors can include a comment line (one starting with a ':' character) every 15 seconds or so.