Analytics: Real Time Error Reports and Client Metrics (Crash Rates, Frame Rates, and Memory Usage)

Hi Creators,

Today, we are excited to launch two new real-time analytics updates to help you improve your experiences:

  1. Error report: View up-to-the-minute errors and warnings for both server and client. Filter by place and severity, search by keyword, and even see stack traces to troubleshoot errors.
  2. Client performance metrics: View up-to-the-minute client metrics (crash rates, frame rates, and memory usage) to spot performance issues early.

You can find your Error Report and Performance pages in the Real-Time section of the Creator Dashboard.

Visit your error report to troubleshoot issues

To view your experience’s error report (e.g., after releasing a new update):

  1. Navigate to your Creator Dashboard and select your experience.
  2. Select Error Report in the Real-Time section.

You can update your error report as follows:

  1. Change the date range
  2. Turn on/off version annotations in the chart
  3. Filter to errors or warnings only
  4. Filter by place
  5. Search by keyword

Based on your updates above, you’ll see a chart that shows up-to-the-minute errors and warnings split by server and client.

Below the chart, you’ll see a table that shows your experience’s errors and warnings by count. You can view stack traces and other details for select errors and warnings by clicking a row that has a down arrow. A stack trace is a list of method calls that your experience was in the middle of when an exception was thrown.

Error reports are available to all creators with published experiences. For more, check out our error report docs.

Monitor your client performance (crash rates, frame rates, and memory usage)

To view your experience’s client metrics:

  1. Navigate to your Creator Dashboard and select your experience.
  2. Select Performance in the Real-Time section.

Scroll down on the page to see your:

  1. Client frame rates: Frames per second processed by each client. Large drops may signal issues.
  2. Client crash rates: Sessions ended in crashes as a percentage of the total sessions ended. Large spikes may signal issues.
  3. Client memory usage: Memory consumed by client. Large spikes may signal issues.
  4. Client memory usage percentage: Percent of available memory consumed by the client relative to available client memory. Large spikes may signal issues.

Note that these charts can be noisy for experiences with low user counts. Your experience needs at least 100 daily active users (corresponds to a few concurrent users) to see these charts.

Troubleshooting client performance issues

If you identify a performance issue, use the following steps to debug and fix it:

  1. Check your recent update and consider rolling it back if the metrics impact is severe.

  2. Use built-in testing tools to troubleshoot performance:

    a. Developer Console for viewing error and log messages and detailed information on memory and networking.
    b. ScriptProfiler: for identifying scripts that take up the most resources.
    c. MicroProfiler for viewing unoptimized portions of your experience visually.

  3. Gather user feedback from your community.

For more, check out our performance page docs.

Build analytics with the community

We’d love to build analytics closely with all of you. If you’d like to preview upcoming analytics features and give feedback, please fill out this quick form and we’ll get back to you.

Please let us know if you have any questions about today’s launch.

Thank you.

390 Likes

This topic was automatically opened after 10 minutes.

Check your calender, it’s not christmas yet!! Well anyways, I’ll take this early present. Thanks so much for this! This is literally so much better than asking 8 year olds what errors they have when they’re character is broken… THANKS!

(edit): How do you prevent exploiters from spamming errors?

129 Likes

I used to spend hours letting my game run for a while to log these sorts of statistics. This is quite the time saver!

Although I would also like to see the most “problematic” tags as well (which memory category uses up the most resources? Which process continuously uses the most resources?). Seeing how it changes overtime is really useful, but giving us pointers to what causes the drain is even better.

84 Likes

This is great, as I won’t have to have an external server logging over thousands of events each day. But how can we figure out who is the exploiter that is spamming false errors, and warnings from their client? Additionally, are there any plans to create a page for custom events?

75 Likes

I believe this is very much needed. So many times I often wonder just how many errors go unnoticed within my games. By finally being able to see the real-time errors that are encountered most frequently, it would be an absolute life saver when testing for bugs and potential errors. I just checked for a few of my games and it looks absolutely amazing, and in my opinion has great potential. This may result in games being less bugged in the future

Overall, this is a great update from Roblox!

65 Likes

Glad to see my favorite team continue to release the most useful updates! Continue to keep up the great work and build products alongside the community!

64 Likes

This is the best update, it will help a lot of developers, thanks for this update

63 Likes

Exactly! Roblox has been making many good updates as of recent, from the VR release to this. Overall, this is like Christmas LOL

59 Likes

Can’t believe this didn’t exist before, this is actually gonna be gold for debugging stuff. Me and the NPCs above are definitely gonna have a great time with this

69 Likes

Holy!! This is absolutely incredible stuff! I will definitely be making use of this to 10x my debugging process.

One feature I would really love to see is the ability to filter errors by Server JobId as well as by UserId, that way, I can hone in on user-specific problems and even user-server specific problems.

Other than that, this is definitely an amazing feature! Bravo!

56 Likes

This will be heavily useful instead of third-party website to track errors and stuff and debug in an almost instant instead of relying on people reporting.

52 Likes

this is great [no more needing to manually log errors to discord] but I wish you could see farther back than 30 days & give a custom timeframe…
this is way too limited
image

along with this, we need:

  • to know who the client log is from
  • to know when the logs are from [timestamp]

it’s not much help if we don’t know who & when

47 Likes

Been waiting for this to be released for months now, and it came at the perfect time! So glad that everyone gets access to this, as there’s new projects I have in private testing, and it will make it so much easier to just open the page up and look at the most recent errors & warnings.

Client crash rate is such a welcome replacement-ish to the old devices which crashed on the old Developer Stats page (which really made little sense),. Would definitely love more information on crashes at some point, like the amount of total memory / memory usage that device had around the time of crash.

Easily some of the most useful analytics features that have been released so far. Thank you so much to the team that made this possible!!

39 Likes

May have found a silent error that’s never been brought to my attention, neat. Also:
image

40 Likes

I wouldn’t want to see errors over a month ago. At that point, those errors could be fixed due to updates and I don’t want those scaring me again.

35 Likes

but say something happens and you cant get to studio to fix all those errors… if you miss them, they’re gone for good.

perhaps they could add a way to mark errors as solved in the dashboard so you don’t have to see old ones?

30 Likes

LETS GOOOO!!!

Finally!!! Now debugging is gonna be so much easier. Big W

Oh…
image

Also, can core script errors not be shown? Literally what am I supposed to do with this

And also, why cant we filter if we want to see only client or server errors? Seems kinda weird to include a severity type, but not where they occured

46 Likes

I am receiving similar errors in my experience that does not include Adonis, it appears as I this may be a falsified client warning fired by an exploiter as it is only occurring heavily on my experiences without any anti-exploit.

I am receiving this specific error a lot (there are more occurrences if I set it to more than 1 day, plus this is a screenshot from one of my smaller experiences)


I don’t have this warning in my code and my experience does not show links nor does it use PolicyService. I can only assume this is a warning being caused by the Roblox Client that is being reported to me as a developer instead? This is unfortunately unhelpful as most other warnings / errors in my experience are caused by the CoreScripts, not me as a developer. There needs to be some way of filtering these out, on top of “audio failed to load” error (I have a UGC based game in-where players can insert audios that may fail to load due to being private in some cases). This error regularly clogs up my error log externally and is now affecting this page.

30 Likes

LETS GOOOOOOOOOOOO! Thank you, Roblox!! This will save loads of time.

28 Likes