All retries have timeout of 1ms set despite not being set anywhere that way

[canpoyrazoglu]

We have the following configuration:

axiosRetry(axiosInstance, { retries: 3 });

However, it never tries to retry. I also tried adding a onRetry handler, putting a console log inside. It never gets hit either, confirming that it indeed never retries.

When I override retryCondition it does get triggered though if I try to call isNetworkOrIdempotentRequestError with the error in the callback, I get Cannot read property 'status' of undefined.

After a failed request, axios-retry is retrying all requests with 1ms timeout, causing all of them to fail repeatedly.

What am I doing wrong?

(React Native 0.71.6, axios 0.27.2, axios-retry 3.3.1)

[yutak23]

it never tries to retry.

Can you provide details of the error when it does not retry so that we can better clarify the issue?

[canpoyrazoglu]

After digging deeper I've realized that the error Cannot read property 'status' of undefined is actually throwing at a place that a colleague wrote long time ago. I've fixed it, however, the issue still continues with another error. First, here is my original request:

{
   "message":"timeout of 2000ms exceeded",
   "name":"AxiosError",
   "config":{
      "transitional":{
         "silentJSONParsing":true,
         "forcedJSONParsing":true,
         "clarifyTimeoutError":false
      },
      "transformRequest":[
         null
      ],
      "transformResponse":[
         null
      ],
      "timeout":1,
      "xsrfCookieName":"XSRF-TOKEN",
      "xsrfHeaderName":"X-XSRF-TOKEN",
      "maxContentLength":-1,
      "maxBodyLength":-1,
      "env":{
         "FormData":null
      },
      "headers":{
         "Accept":"application/json, text/plain, */*",
         "user-device-id":"[redacted]",
         "x-client-version":"1.0.17",
         "Authorization":"Bearer [redacted]",
         "x-locale":"es"
      },
      "baseURL":"https://[redacted]",
      "method":"get",
      "url":"[redacted]",
      "axios-retry":{
         "retryCount":1,
         "lastRequestTime":1693897894950
      }
   },
   "code":"ECONNABORTED",
   "status":null
}

Notice that it has a timeout of 2000ms (normally it's 12 seconds though I reduced it temporarily for testing this issue). After the requests fail, I get this:

timeout of 1ms exceeded

Why is it trying again with a 1ms timeout? After initial (failing) request, it will (obviously) always fail again as timeout is 1ms. I think exponential backoff would be more appropriate.

My observation is that after initial (real request, made by me, not axios-retry) the timeout is set correctly at what I've provided. However, all the subsequent requests (the retry requests made by axios-retry) have a timeout set to 1ms, failing in rapid succession.

I'm not setting timeout anywhere to 1ms. THe only reference to timeout (either in axios or axios-retry in total) is only where I create axios instance:

Axios.create({
  baseURL: Config.API_URL,
  timeout: 2000,
  headers: {
    [...]
  },
});

So even if the error in my original issue was misleading, the actual issue remains. (I've updated the original issue and title to reflect the actual problem)

[mishagale]

I suspect this might be caused by not setting the shouldResetTimeout flag to true.