Project usage tab doesn’t seem to work for me using the simple dashboard display.

Displays all $0.00

Oops, was only enabled for admins. Should be fixed in a few minutes. 👍

Works, thank you!

Wauw nice job, looks very good!

Nice improvement! To me, it would be very useful to be able to set alerts based on user-defined thresholds.For instance, I might want to be notified via mail if my services exceed 2 GBs of memory usage

Can you give an instance one color to be always the same? CPU usage shows a different color code for an instance than memory usage.

It's looking great so far! Some initial feedback:- The smallest chart vertical size is huge! - In general, I would love more compact widgets. They have a lot of padding and the header is very large- Hovering the heading makes the content jump to show the extra description- It's frustrating to have to grab the little "move handle"- Would be nice to have every corner/edge be able to drag to resize- "Provide Feedback" button covers the drag handle of bottom-right widget(Edit: Looks like I lost all my formatting in this comment)

The forum now uses tip tap for format editing, so there are some bugs to work out still haha

Nice improvement! To me, it would be very useful to be able to set alerts based on user-defined thresholds.For instance, I might want to be notified via mail if my services exceed 2 GBs of memory usage

For sure. It's a little bit away but the eventual plan here is to left you define alerts on top of the widgets you create in these dashboards!

Can you give an instance one color to be always the same? CPU usage shows a different color code for an instance than memory usage.

I am going to play around a bit with colors either this week or the next. There's some nuance to work out though:

• After a certain number of services, the colors can't really be unique and there will be some form of similarity/collision between services across charts, I wonder how much having the same color helps at that point.

• On a similar note as above, if you have a lot of services, would you really use color to identify them or would labels and filters make more sense?

I don't have a large amount of services right now, hence I would try to identify them by color. If I check the CPU usage chart and my service X is green, I expect it to be green on the memory and disk chart as well.

The smallest chart vertical size is huge!

I went back and forth on minH but for some reason after reason this, it seems obviously huge, haha! IIRC it's 3 units right now, I'll probably bump it down to 2 and take it from there.

Hovering the heading makes the content jump to show the extra description

This is a bug. Will fix! (And will also look into the more compact charts feedback).

It's frustrating to have to grab the little "move handle"- Would be nice to have every corner/edge be able to drag to resize

Hmm, so this was intentional because it felt too noisy otherwise? Originally, you could just drag the whole card around. But, we didn't really have edit mode back then so maybe that's why it felt noisy. I'll put this down as a wait and see in terms of how many people complain about it. :P

"Provide Feedback" button covers the drag handle of bottom-right widget

Will fix!

I don't have a large amount of services right now, hence I would try to identify them by color. If I check the CPU usage chart and my service X is green, I expect it to be green on the memory and disk chart as well.

FWIW this will be the short term solution. I guess what I was looking for with that comment was whether you'd use the same notion for identification if you did have a lot of services or would you prefer some form of global filtering mechanism?

Would use the same method, additional to labels.

Small UX feedback: - "Add item" should be "Save item" or something similar when updating instead of creating new- Sometimes you can't move resized panel above the not-resized one, you have to resize both only then it works

Update:

• Cards can now have a smaller height than before.

• Feedback button doesn't cover the resize trigger on the bottom-most widget.

• Hovering to view description doesn't cause any layout shift.

I don't have a large amount of services right now, hence I would try to identify them by color. If I check the CPU usage chart and my service X is green, I expect it to be green on the memory and disk chart as well.

This is done. Your resources will now re-use the same color across different widgets!

Nice. Thank you!

Just a thought: I would find helpful to be able to set custom colors (even defined by their hex code?) for each service. Since we might associate them with a specific color (based on their logo, etc), this can improve dashboards' immediacy.

Hello I was wondering if we can get a "count" or "total" on this particular tile. Reason for this is it'll make it a lot more valuable. It'll show logs but gives the user an opportunity to configure logs to show many times an endpoint has been hit for example or what is the error ratio for an endpoint or even the error rate across the board.

It's very strange to me that the Railway team implemented this kind of dashboard instead of just adding log drain functionality.

If your goal is to do everything in-house, we NEED alerts. Your serious customers must be able to trigger alerts (email, Slack, etc) based on application output.

It seems like Railway wants to build an observability product similar to BetterStack, which isn't ideal but would at least get the job done.

I'm a huge fan of this new feature. This would be unbelievably useful to me if I had the ability to embed this view in an Iframe - I'd even pay extra for this feature. Embedding these logs directly in another dashboard would be incredibly helpful. Are there any plans to generate sharable versions of the dashboard a la Plausible?

Thanks

We agree that would be a cool feature to have but it's just not possible right now, all that data uses your login token, and so the embed would need access to that too, this would only be do-able if we introduced some finer grained permission controls, and that's unfortunately not something we could do right now, but thank you for the feedback!!

I love this feature. Expect to upgrade to Pro soon, just testing out the platform for now.

As for this feature, I'd love if there were "templates" for different use cases, resizing of windows, and multiple "pages" or "tabs" so that I can create one for project usage, one for errors, etc.

All the best!

I love the feature, but I think it would be even better if there was an option to create tabs or sections. I’ve deployed 3 services in my project and created individual cards for each, along with a common card for all services. So, I have to scroll quite a bit to view a specific service graph.

Also, it would be really useful if there were an option to resize the cards.

Hello, you can resize the cards, click the Edit button and then use the resize handle to drag to the size you want! once done, click Save.

I've been really enjoying the observability dashboard but it's almost impossible for me to rearrange the panels. They seem to have a mind of their own and move around quite a bit after saving a layout.

(I tried to attach a video to this thread but it doesn't seem to be uploading, you can contact me on discord if you do want a video )

Can you open a #feedback thread in our discord server so you can share the video?

https://discord.gg/railway

Love the observability dashboard! Is there any chance we'll get HTTP Logs in there any time soon?

This is an awesome feature — great work, team!

One thing that would make it even more powerful is support for alerting and log drains. Until building in-house alerting and analytics, it would be super helpful to have built-in ways to trigger alerts (like email or Slack) based on log output.

I'm sure a lot of teams would benefit from this kind of observability — especially during critical production deployments or error spikes.

Would love to see this added!

Yeah. We still don't have Slack alerts based on log output. It's really quite sad the Railway team hasn't prioritised this.

We have to resort to service-internal workarounds to actually pipe our logs somewhere that can do alerting, like Papertrail.

Alerting based on (structured) log output is table stakes for any prod service.

Perfect timing—our team is kicking off work this quarter to bring custom alerts to Railway!

We’d love to hear how you would use alerts and what you’d expect from a built-in solution. Keep an eye on our Public Roadmap for updates, and feel free to drop any comments or feature requests there—we’re all ears!

We totally hear you and we're committed to changing that this quarter! The team is beginning the work to bring alerting to Railway and we hop you'll join us for the ride as we work on it.

We’d love to hear how you would use alerts and what you’d expect from a built-in solution (Slack alerts is a good one, we'll definitely keep that in mind).

Keep an eye on our Public Roadmap for updates, and feel free to drop any comments or feature requests there—we’re all ears!

We’d love to hear how you would use alerts and what you’d expect from a built-in solution

Well, it's disappointing to hear you imply you will not be adding industry-standard log drain functionality... If you truly want to keep everything in-house, yes – Slack alerts based on any arbitrary log query is what is needed by our team, specifically.

The new observability dashboard is nice, one piece of feedback that would really help our team is granular breakdowns of the network egress. Currently the network egress just shows egress in aggregate without any information of where or what that egress is attributed to. If we could see a breakdown such as the domains or individual requests which are adding to the network egress, that would help a lot in identifying areas to optimize and improve within our application. Not sure if this is possible to do from Railway, but if it is it would be tremendously helpful.

The vertical bar should be displayed on ALL graphs, not just the one I'm mousing over. This would help tremendously with visual correlation of events across multiple resources.