Skip to tool
TRANSPORT DIAGNOSTICS · CONTROLLED ENDPOINT

SSE stream stability

Validate six SSE events continuously or across one planned interruption and explicit Last-Event-ID resume.

Checking network test availability…

Check event continuity and controlled recovery

This runs against our fixed endpoint. It uses a fetch-based SSE reader with signed session headers, not the browser’s native EventSource client. No measurement starts on page load.

The server sends events 1–3 and a checkpoint, then deliberately ends the stream. This client waits 500 ms and resumes once with Last-Event-ID: 3, expecting events 4–6.

Data events
6, in sequence
Nominal cadence
250 ms
Response deadline
6 seconds
Run deadline
20 seconds, plus cleanup

At most two streams, each with a 32 KiB read cap. Admission reserves an 8 MiB allowance without sending it; each stream debits 256 KiB conservatively. Stop also requests server-side stream termination.

Interpret the result within its scope

Sequence validation checks this diagnostic’s data contract, not an arbitrary application. A continuous run does not test reconnection. A resumed run tests one deliberate checkpoint and one client-managed request, not retry storms or every disconnect condition.

Browser scheduling, server handling and intermediate buffering can change arrival gaps. Neither gaps nor failed streams are IP packet-loss measurements. No Wi-Fi, router or ISP cause is inferred.

Methodology and limitations · Transport tools · All tools