Announcing Observability Dashboard for HttpService

Updates Announcements
1

Hi Creators,

Today we are announcing the launch of the HttpService Observability dashboard. This new page on the Creator Dashboard provides real-time visibility into your Experiences’ HttpService usage, allowing you to monitor request performance and debug issues more effectively.

The HttpService Observability dashboard has 2 charts: Request Count and Response Time.

  • Request Count tracks the number of HttpService requests from your Experience.
  • Response Time measures the latency of your requests, or how long it took for the endpoints to respond.

The dashboard includes additional controls for filtering and breaking down your data by several key dimensions:

  • Filter by Place: Isolate metrics to a specific Place.
  • Breakdown and Filter by Request Type: See your request volume and performance segmented by the designated HTTP methods.
  • Breakdown and Filter by Response Status: Analyze your requests based on their response status to quickly find errors.

image2
image21407×915 76.3 KB

image1
image11361×953 91.9 KB

How to Get Started

You can access this new dashboard by navigating to your Creator Dashboard, selecting your Experience, and selecting Http Service under the Monitoring section. For further information on the dashboard, take a look at the Http Service Observability documentation.

For future releases, we’re considering additional breakdowns such as by HTTP domain. If this would be useful for you, please let us know more about your use case for this control. As always, please let us know if you have any ideas for charts or breakdowns and why they’d be valuable to you. We look forward to your feedback as we continue to improve the HttpService and Observability tools.

112 Likes
2

This topic was automatically opened after 11 minutes.

3

This would be very cool and useful if the platform itself wasn’t ruined.

Thanks for this useful tool either way, if you every fix the issue that I don’t need to mention, I will find much usefulness in this tool.

39 Likes
4

Would be cool if you could break it down by URL to see where the calls are going

19 Likes
5

Yes and the exactly unencoded headers, api keys, and options

3 Likes
7

Just asking for when this’ll be possible in the future; for SSE connections, will it use time to headers rather than time to end for response time?

3 Likes
8

I really like all these HttpService changes, it gives me hope WebStreamClients will be enabled in RCC soon.

3 Likes
9

We got observability dashboard before gta 6

1 Like
10

I didn’t even know I wanted this. :o

It would be nice to see an overview of all of the URLs that the experience is hitting and how often as well, it would make it more clear when something unusual is going on, particularly with large un-auditable third party code.

4 Likes
11

This is cool. It’s now possible to see how much it can be used and to find ways to regulate the use of HTTPService.

1 Like
12

Hi Scripty,
Thanks for sharing your interest in this capability, we are looking into how we can support domains, given the large number of variations in URL, but would love to hear more about how it would benefit your experiences!

4 Likes
13

I think domains would be a good first step.

1 Like
14

What is the grounds for the hate? I’m confused as to what is ruining the platform

3 Likes
15

It’s a good way of identifying something that shouldn’t be happening, for example if I build a system that is supposed to make a Http call to a specific domain 1 time per minute, but then in the analytics I see it’s doing it 100 times per minute, then I can easily identify it.

16

Please OH PLEASE add a firewall to unauthorized links so we don’t have to deal with backdoors.

Or at least add a studio mode. I’m tired of dealing with proxy metatables to monitor HttpService in obfuscated scripts.

17

why are you using obfuscated code in the first place??

8 Likes
19

It should be obvious by now that this is concerning age segregation.

5 Likes
20

Posting about it on every new annoucement won’t fix anything and is be annoying though

1 Like
21

I’m simply answering their question by clarifying what they meant, nothing more.

3 Likes
22

Breakdown by domain would be great, but to really make this useful, it’s important to know the status code of these responses too – latency & quantity alone is not particularly useful. A dedicated always-present chart showing status codes, or perhaps even a pie chart for an instant in time of status codes, would both be great. This is neat right now as-is-, but not very practical to use.

It’d also be nice to be able to breakdown by data centre / region origin, so we can isolate latency issues + error rates to specific regions.

Additionally, all of my error codes are currently HttpService InternalError, which is pretty much useless for me to try to figure out what the issue/cause is there. More specific internal error codes would be great.

4 Likes