Intercept timeout errors?

I use interceptors.response.use to log details about the for every request we do (error or success). It’s very useful to have it in our logs to investigate on issues. But when we get a timeout error, those interceptors are not invoked so all we get in our logs is:Error: timeout of 3000ms exceeded Is there some kind of global event or something that I could use to add context to timeout errors?

Context

• axios version: v0.16.2
• Environment: node v8.5.0

[nickknissen]

When a timeout happens it will return an error code with the value ECONNABORTED https://github.com/axios/axios/blob/b7d8d126dad7341addcd7b063f7b261955ec9152/lib/adapters/xhr.js#L95

This is more or less how i do it:

axios.interceptors.response.use(
  config => config,
  (error) => {
    if (error.response.status === 408 || error.code === 'ECONNABORTED') {
      console.log(`A timeout happend on url ${error.config.url}`)
    }
    return Promise.reject(error);
  },
);

[davidlukerice]

The ECONNABORTED code is also used when the request is manually aborted: https://github.com/axios/axios/blob/503418718f669fcc674719fd862b355605d7b41f/lib/adapters/xhr.js#L71 Is there a good way to differentiate between the server request timeouts and manual aborts?

We’re not currently using .abort() anywhere, so that check works for us for now, but I could see a bug crop up if we later had to add in a manual request cancel and forgot about our assumption of using that code only for the timeout case.