Request for feedback - UI dashboard

We’re planning to redesign the Mender UI dashboard a bit to improve the quality and scalability of the information displayed on it. A lot of the current dashboard data is duplicating what’s in the deployments tab and it could be a lot better.

I would really appreciate some feedback from Mender users about what you’d like to see on the dashboard, what information is important/urgent for you to know every time you log in to the UI. Some suggestions:

  • If there are pending devices
  • If the number of accepted devices has changed recently - have any been accepted or removed since your last login?
  • If any deployments have completed recently/since your last visit
  • If there are deployments in progress
  • If there are deployments pending
    • Q: how long can a deployment stay in ‘pending’ status before you start to think something is wrong?
  • Show status of your most recent deployments - for example the status all created today & yesterday (and even longer back? 3 days?)

Do these sound useful, is anything missing? What do you look for first when you use the Mender UI? What would events would you want to be notified about?

Your comments would be very valuable!

1 Like

Your suggestions sound useful. I’m not sure about the actual implementation but I think it might be useful to have a quick overview on the dashboard of the number of devices that have heartbeated recently and the number of devices that haven’t heartbeated in X amount of time.


Currently, all the devices are listing on the server based on either the device ID (which server assigned) or the MAC ID (specifically Ethernet MAC ID).

There are some cases, where the device might not have an Ethernet connection (it may have a Wi-Fi or LTE connection instead).
We know how to get the required information by changing the script on the device.
But, I would like to know whether any custom option can be given on the server (apart from the MAC ID device ID) to name it as what we required?

@WeZzyNL that’s a good suggestion. I would also like to show that on the dashboard - something like “X devices may be offline” if they haven’t heartbeated in user-selected amount of time. And then click through to be taken to view those devices. What sort of timeframe would be a sensible default in your opinion?


Not on the server as it stands, but you are able to configure the device identity on the client - - but I think that might be what you are referring to.

This topic was more about the dashboard tab on the GUI, the first page you see when you load it up - if there was some information you would want to see on that page to get an at-a-glance overview of your devices/deployments, before you dig in to details.

1 Like

First, I would like to thank you for a great product. I agree with the above folks, a status of connected clients by group would be No. 1.

The rest would be nice to help with problem clients…

Average time to complete and verify successful updates. We have devices that turn off at night. It would be nice to know the latest time we can update based on the average update time.

No. of retry attempts for updates if greater than one. – Again show’s problem locations as we are geographically spread. Showing if there are issue in smaller community where bandwidth is limited.

No. of successful updates for the month by group.

Bandwidth usage especially during updating.

1 Like

@gil0109 Thanks! That’s very helpful.

We’re about to add a basic feature on the dashboard that alerts you when clients haven’t pinged in 24 hrs - we will iterate over this and make the length of time customizable too.

Some of other features you suggest sound super interesting, and might mesh well with an idea for per-device update history that we have in mind.

(On a side note, if bandwidth usage is an issue for you then Update Modules might be interesting to you!)

1 Like

i am not aware of an option in the dashboard to do an deployment at a certain time or at least did not find it, but this would be nice to lets say for example 01:00 am push some deployments to devices.
is this possible at all? with or without the dasboard?
anyway this would be a nice option to have :slight_smile:

They way you would implement something like this today is on the client utilizing state-scripts which allow you control the update flow, here you can say, only start downloading the artifact if it is after 01:00, or you can download it but do not reboot until time is 01:00 and so on.

1 Like

I personally would like the ability to group devices (for both monitoring and releases) by one of Inventory fields, and make these groups dynamic (as inventory is). I understand it is possible by using API but having this out-of-the box would be cool!

And a couple of comments regarding today’s UI:

  • Deployments view shows deployments of the last day by default and no quick way to show all deployments.
  • impossible to create a deployment for a number if individual devices

As for proposition for dashboard, It would be nice to have a system overview where you can see the percentage of all software verions deployed in your system

Firstly, I want to thank you for a great upgrade in v2. I am currently running it in OpenShift (that was fun to get working) and so far things are working great.

Suggestions for UI now that I am using the new GUI for a month or so:

Display version of Mender
Ability to see offline by Group
Ability to see only offline devices by clicking warning
In Device area, it would be nice to sort by columns displayed (especially Last Checked in)
In Device area, Ability to search for a specific device by various methods mac/location/etc…

I haven’t done a large push of an image update yet but that’s coming…

Hi @ster and @gil0109,

Thanks a lot for your feedback, this helps make Mender better for everyone!

Here are some comments:

  • Showing all deployments. This used to be the default, but due to scale issues of loading everything by default we switched to last day default (it got slow for some larger installations with 10s-100s of deployments each day). But maybe last week would be a more reasonable default? @michaelatmender?

  • Create deployment for several individual devices. This is what groups are for. Or did you not want to create a group for these devices – are there any issues you see with this?

  • Overview of software version distribution. This is on the roadmap and planned as a commercial feature (Hosted Mender / Mender Enterprise).

  • Display version of Mender. This is an inventory attribute and should be available by default. Can you click on a device under Devices and see if you see the field “mender_client_version”?

  • Offline by group. Good point, will consider for new feature (fits together with software distribution).

  • See offline devices. Same as above, I think this is part of the same feature (not very useful to see how many devices are offline if you cannot see which ones).

  • Sort by columns. On the roadmap for all editions.

  • Search for specific device. Filtering by inventory attributes is already available in commercial editions of Mender (Hosted Mender / Mender Enterprise).

Overall the goal for commercial editions (Hosted Mender / Mender Enterprise) is to support larger-scale environments and higher levels of security. The open source edition will have the important features and will of course continue to be improved as well!

Thanks again for the feedback, and please let us know if you have more.

The Mender server version is not available via the API, nor the UI.

Hi @SuicidalLabRat,
The Mender server version should be on the bottom left in the UI (at all times), I believe it was introduced in Mender 1.6. Can you not see it?

Sadly, no. We are running 1.7.1; see attached.

That is strange! I just brought up an 1.7.1 server following the docs and it does state the version – see attached.
Are you sure you’re running 1.7.1 and all the services?

Perhaps @michaelatmender would know in which cases this is not displayed?Screenshot%20from%202019-08-06%2010-42-24

It could be that the ‘version’ environment variable isn’t available to the UI for some reason, but I’m not sure why that would happen. The file gets the version from the integration service and inserts it into the UI code. Maybe check if your docker-compose or demo script have been changed and check that there is a reference to INTEGRATION_VERSION?

It doesn’t look like the version is exposed in the production environment script, only in the demo environment. So it will not currently be seen in production environments.

This is a bug, I have created

Quick FYI
We have upgraded to 2.0.1 and this issue persists. i.e. no version number displayed in the UI.