Retry handler should retry on connect timeout

[Sytten]

Feature Request

Is your feature request related to a problem? Please describe

Currently the retry handler only rethrows when the underlining request throws:

msgraph-sdk-javascript/src/middleware/RetryHandler.ts

Lines 176 to 189 in 130f3c0

	try {
	await this.nextMiddleware.execute(context);
	if (retryAttempts < options.maxRetries && this.isRetry(context.response) && this.isBuffered(context.request, context.options) && options.shouldRetry(options.delay, retryAttempts, context.request, context.options, context.response)) {
	++retryAttempts;
	setRequestHeader(context.request, context.options, RetryHandler.RETRY_ATTEMPT_HEADER, retryAttempts.toString());
	const delay = this.getDelay(context.response, retryAttempts, options.delay);
	await this.sleep(delay);
	return await this.executeWithRetry(context, retryAttempts, options);
	} else {
	return;
	}
	} catch (error) {
	throw error;
	}

This means that a connect timeout will not be retried. This happened in production for us the other day.

{ statusCode: -1, code: "FetchError", message: "request to https://graph.microsoft.com/v1.0/me/mailboxSettings failed, reason: connect ETIMEDOUT 40.126.6.114:443", requestId: null, date: {}, body: "FetchError: request to https://graph.microsoft.com/v1.0/me/mailboxSettings failed, reason: connect ETIMEDOUT 40.126.6.114:443" } 

Describe the solution you'd like

The retry middleware should catch the connection timeout and retry.

AB#6270

[melya]

The world is imperfect, and so is the network. A retry mechanism should be implemented for failed requests due to network issues.
Any updates on this?

I've been facing this issue more frequently lately