What timeout should the server use for a long poll?

Below the shortest idle timeout of anything between the client and your server. AWS ALBs default to 60 seconds and Nginx's proxy_read_timeout defaults to 60, so 25 to 30 seconds is the common safe choice. On timeout, return an empty 204 and let the client immediately reconnect; the timeout response doubles as proof the connection is still alive end to end.

How is long polling different from SSE?

A long poll delivers one response per request, then the client reconnects, while SSE keeps a single response open and streams many events down it. SSE also ships reconnection and resume via Last-Event-ID in the browser's EventSource API, which long polling makes you build yourself with cursors. If the server can push and traffic is one-way, SSE is usually less code for a better result.

Do I lose messages between polls?

You will unless you design against it. Events that fire in the 50 to 200 ms gap between one response and the next request have no parked request to ride. Carry a cursor: each response includes the last delivered event id, the client echoes it back as since= , and the server replays anything newer before parking again.

Real-Timechapter 2 of 4 · 3 lessons

Long Polling

Server holds the connection open until data is ready, then client immediately reconnects.

← All chapters View in cheatsheet

Your chat app asks the server every three seconds whether anything is new. Almost every time, the answer is no.

At 10,000 people online that is 3,300 requests a second, and 99 percent of them return an empty list. You are paying for a fleet of servers whose main job is saying nothing yet, and a message still takes up to three seconds to appear.

Long polling changes one word of the deal. When the server has nothing to say, it says nothing and holds the request open instead of answering.

The message now arrives the instant it exists, and a quiet client makes two requests a minute rather than twenty. Your cost stops tracking the clock and starts tracking what actually happens.

Lessons

3 in this chapter

The Hanging RequestThe server parks the request until there is something worth answering with.2 min
Timeouts and ReconnectsYour hold time is set by the grumpiest proxy in the path, and the reconnect gap eats messages.2 min
Long Polling at ScaleCapacity is measured in parked connections, and the failure mode is the reconnect stampede.2 min