Currently, there is a signal parameter which accepts an AbortSignal but this has two limitations which make configuring request timeout not feasible:

1. If a request gets timeout, the signal.aborted field becomes true, and it stays true in subsequent requests as the object remains the same.
2. When we get a timeout, in the request method of HttpClient, a Response object is created with status 200. So the request method never throws an error. The response is simply null.

Did the following to fix the issue:

• Made it possible to pass an AbortSignal creator function as signal.
• When this is done, we simply throw the error which can be cought like this:

const redis = new Redis({
  url: "***",
  token: "***",
  signal: () => AbortSignal.timeout(500),
});

try {
  const count = await redis.hset("myhash", {
    key: null,
  });
} catch (error) {
  if (error instanceof Error && error.name === "TimeoutError") {
    console.error("Timeout: It took more than 500 ms to get the result!");
  }
}