End User Monitoring | Datadog

End User Monitoring

Published: July 17, 2019

A few months ago, we launched Datadog Synthetics.

It was our first step in the world of monitoring end-user experience.

Since 2010, the goal of Datadog has always been to break down silos to ensure that development and operations team could work together easily.

We are now moving up the stack and finding ways to give you a better picture of performance from the perspective of the final users of your products.

A study by Google shows, for example, that Pinterest could increase signups by 15% by decreasing the perceived wait time by 40%.

On the other hand, the BBC found they lost an additional 10% of users for every additional second their site took to load.

And this is why we care about digital experience management.

We currently allow you to test your user journeys and endpoints from 13 public locations around the globe.

But what if you’re looking to monitor internal applications or from a very specific location like a remote office?

Private Locations

Starting today, private locations will allow you to set up a new synthetic test region in less than five minutes and have them displayed just like all the other locations we provide.

You can then test any endpoint and any user journey even if they are not on the public internet.

For instance, we could set up a private location right here in New York to alert our engineering team should end-user experience, from the Dash expo floor right over there, not meet our SLAs.

But imagine all of a sudden, one of the critical features we were to demo today stops working.

For example, imagine the UI allowing users to re-index logs from archives stops working.

You are alerted by your failing synthetic tests, but now what?

Remember, back in the day, whenever an end-to-end test, whenever a user journey and prediction stopped working, you had to reproduce the issue, then open your console to search for frontend errors,

understand the issue comes from the backend, and then investigate every piece of software involved in it to understand where the issue comes from.

Well, now, you can go much faster.

Synthetics and APM integration

Starting today, alerts from Synthetics will automatically come with the relevant traces and logs, allowing you to skip all of these steps and get straight to the root cause.

For every request made during the navigation of a browser test and every request made during an API test, we will display the backend trace triggered.

You not only get alerted about a symptom of end-user pain, but also with all the contexts you will need to troubleshoot and identify the root cause, whether it be a problem on your frontend, backend API, or even a network connectivity issue between you and your users.

But let’s see that in action.

This is a browser test that we have recorded to ensure that all the time, the signup to Datadog and the main page of the synthetic app work correctly.

And here you can see the test is failing.

First, I will review the screenshots to understand what happened to the script, and I notice the error.

Then I will check my frontend errors, and I find a 502 error, meaning from the frontend side, we are blind, and the issue is on the backend side.

We can then jump to the exact backend trace associated with it and identify the related error, in that case, coming from undefined value handled in this backend service.

Real User Monitoring

When you start working on the optimization of your applications, you need to know more.

You need information about the latencies in the real world and the number of users experiencing them.

This is a lot of data to collect and analyze, but else, how could you prioritize your work?

I am glad to present you the solution because our brand new product, Real User Monitoring, can help you with that.

Real User Monitoring collects telemetry on every user session on your website and web application.

Tightly linked to Datadog APM, you will be able to see traces of frontend performance alongside backend traces.

It means, if the performance of your products are not good enough, you can understand at one glance whether the issues come from the frontend or the backend.

By seeing the differences across broader versions, geography, and devices, you can investigate the performance issues and make sure your product stays on top of your users’ expectations.

Using information about the real usage of your product, you can prioritize your actions to improve the performance of the most visited pages first.

Real User Monitoring is in beta, starting today.

Moving up the stack is an exciting journey.

We believe in the added value of offering you products contributing directly in the improvement of the experience of your customers.