I would suggest (if it is going to be accepted at all) to make it an option in the configuration file, not in the UI itself.

The problem with UI (all of them) than literally nothing could be configured in the configuration file - everyone gets the same set of defaults and all changes are only stored in a browser (or cloud) while not everyone is using one browser or allows cookies and not everyone is using cloud.

thanks for sharing your feedback @foxt will ask some further question for us to try to understand from the old UI what you're really missing

Currently you havbe to have click in the area where the old link is when linking and hope you clicked in the right spot and clicked fast enough.

the old UI is still available, if you go to /v1 you can get to old dashboard - like this: https://sanfrancisco.my-netdata.io/v1

There are also concerns about Netdata automatically installing propriatary code licensed under a new UI, and loading code from the Netdata servers (which may be changed at any time or monitored for analytics) (see also, #15466 (comment))
The link for the license is only shown in a very small link for a brief period of time (only 1 second)

there was a big concern with the licensing around the new UI and all efforts were made to ensure we keep the Netdata Agent as open source but for sure we'll further investigate this concerns

regarding the link for the license you can find it here - https://github.com/netdata/netdata/blob/master/web/gui/v2/LICENSE.md

Value proposition
The reason I ask is the new UI makes the assumption that you

1. are using multiple nodes with ND installed.

could explain a bit why for the case of monitoring a single node the new UI doesn't meet your needs? is it the multi-instance charts (discussed on a discussion mentioned below) or something else?

2. are using the cloud hosted service.

the new UI is totally independent of you having a Cloud account or not, there are some features related to authorization/authentication (like Functions) or other related with ease of customization that link to a Cloud account but simple monitoring and troubleshooting like it was on the old UI should still be possible. if you find something amiss could you share with us?

For example, if you are not logged into the cloud, you cannot change any Netdata settings in the new UI.

it is true that access to the old UI settings is no longer available but most of them should be available on the new UI directly on the charts or other components. if there is a setting that you find missing do let us know and we'll try to see if can bring it into the specific components.

also there is this GH discussion about this move to the new UI where we explain the rational for the move and also confirm that we need to work on an expanded view to give back instance charts (not sure this is what you mostly miss)

thanks for the suggestions @aldem , will try to further explore some of them to better understand your needs

I would suggest (if it is going to be accepted at all) to make it an option in the configuration file, not in the UI itself.

this indeed could be a configuration, we'll check this internally

The problem with UI (all of them) than literally nothing could be configured in the configuration file - everyone gets the same set of defaults and all changes are only stored in a browser (or cloud) while not everyone is using one browser or allows cookies and not everyone is using cloud.

could you share what configurations you would like to see available in the configuration file?

the new UI is totally independent of you having a Cloud account or not
..
it is true that access to the old UI settings is no longer available but most of them should be available on the new UI directly on the charts or other components.

This is not entirely true. The problem with new UI is that is has too many tunables and any change that you make in any chart (grouping, aggregation, chart type etc) is not preserved - once you close the browser or reload the page everything is lost.

Some settings (like metrics in "Nodes" tab) are not preserved across reload too, you get a message "Connect to cloud to be able to save your settings".

While v1 dashboard does not have many of those settings, it does show more relevant information by default even without customization.

Yes, v2 is very nice, very advanced and useful, it is just... not really user friendly by default in some parts, especially for those of us who is roaming between browsers too often, or who does not have dashboard open 24/7 (which is difficult because of enormous RAM usage if you leave it open & focused for a while).

could you share what configurations you would like to see available in the configuration file?

• default dashboard's version
• timeframe
• for every chart (by context) - default grouping mode, aggregation and chart type, as existing defaults sometimes quite meaningless (like summing up latencies or displaying average across significantly different ranges)
• filters for events adn alerts (to show only "Errors" and "Critical" for instance)
• metrics displayed in "Nodes" tab

The problem with new UI is that is has too many tunables and any change that you make in any chart (grouping, aggregation, chart type etc) is not preserved - once you close the browser or reload the page everything is lost.

agree - just cross linking this feature request that might help: netdata/netdata-cloud#789

This is not entirely true. The problem with new UI is that is has too many tunables and any change that you make in any chart (grouping, aggregation, chart type etc) is not preserved - once you close the browser or reload the page everything is lost.

Some settings (like metrics in "Nodes" tab) are not preserved across reload too, you get a message "Connect to cloud to be able to save your settings".

I was comparing with the old UI settings, e.g. Dark theme (we have some bug we will introduce this again) and the time, the configurations per chart (grouping, aggregation, chart type) is something new that you could couldn't manage on the old UI

While v1 dashboard does not have many of those settings, it does show more relevant information by default even without customization.

this is really what we are trying to understand, what from v1 dashboard is missing - we understand some way to store and save chart defaults and Node tab metrics is desired.

could you share what configurations you would like to see available in the configuration file?

default dashboard's version
timeframe
for every chart (by context) - default grouping mode, aggregation and chart type, as existing defaults sometimes quite meaningless (like summing up latencies or > > > displaying average across significantly different ranges)
filters for events adn alerts (to show only "Errors" and "Critical" for instance)
metrics displayed in "Nodes" tab

thanks for the share a bit more on what you mean by:

• timeframe?
• filters for events adn alerts (to show only "Errors" and "Critical" for instance) --> do you mean default selected filters?

@hugovalente-pm

the configurations per chart (grouping, aggregation, chart type) is something new that you could couldn't manage on the old UI

Yes, I couldn't - but I don't need to as everything is visible, I just have to scroll (or click menu on the right). In v2 I have no choice and must click more times to choose what I need, every time - even after I switched to Events and then back to Overview, and this is really annoying.

By timeframe I mean the default period covered when I open the UI - last 10 minutes, last day etc.

Filters for events and alerts - yes, I mean default selected filters.

even after I switched to Events and then back to Overview, and this is really annoying.

This is a bug, we previously supported this but somehow on the process of the merge of these UIs it was broken - there is this bug to track it netdata/netdata-cloud#903

By timeframe I mean the default period covered when I open the UI - last 10 minutes, last day etc.
Filters for events and alerts - yes, I mean default selected filters.

These are some interesting visualization preferences to add, we don't even support these on Netdata Cloud. Made this feature request to track this netdata/netdata-cloud#904

the old UI is still available, if you go to /v1 you can get to old dashboard

Is it still receiving bug fixes and security fixes, if yes for how long ?

Thanks

@KaKi87 the old UI is held on netdata/dashboard if you check the README there it mentions this is deprecated

So the answer is no, so using it is not an answer to this issue.

So the answer is no, so using it is not an answer to this issue.

The path forward is improve and make needed changes on the new UI, in the meantime the old UI can be used but no work will be done on it

Well, the improvement we want is basically a KISS mode : Keep It Simple Stupid.

An uncomplicated mode.

Thanks

The path forward is improve and make needed changes on the new UI, in the meantime the old UI can be used but no work will be done on it

This means that there is no longer a supported open source UI. Which makes your claim invalid. @ktsaou https://www.netdata.cloud/open-source/

Please communicate this clearly and inform the users about the proprietary frontend license. Or rather make the frontend open source!
I assume that most users are not aware that the UI is no longer Open Source.

@hugovalente-pm

So the answer is no, so using it is not an answer to this issue.

The path forward is improve and make needed changes on the new UI, in the meantime the old UI can be used but no work will be done on it

How would you like to go about this? I think a lot of users have commented on the negative value to aggregate different metrics as one.

Some examples
IPMI monitoring combines fan speeds as one metric. This is counter-productive as faulty fans aren't visible. I simply can't see if one of 6 fans is running slower than usual. The same applies to temperatures and voltages.

In a similar way I cannot easily draw conclusions out of any of the container/cgroup charts as averages do not really give insight in the health of individual services. Even on a relatively small system, a service that goes rogue and pins one CPU would not be visible unless I am actively looking for that service by filtering the chart on the service.

In general, I'd say this is the major drawback of the new design. The possibility to have a quick glance to see out of order issues/stats is removed. Netdata now provides an overview, but few details, which is at odds with the enormous capacity to gather detailed statistics.

How would you like to go about this? I think a lot of users have commented on the negative value to aggregate different metrics as one.

We have two main tickets that relate to what we plan to work on this:

• [Feat]: ability to "expand by" or similar for charts (e.g. expand by container, mount point etc) netdata-cloud#894 --> allow the expanding of charts based on some attribute, for example it could be the instance
• [Feat]: ability to save/override chart defaults in a Space or Room netdata-cloud#789 --> allow the storing of chart settings

We are also working on improving the Custom Dashboards that will allow to make your own dashboards in an easier way

We weren't able to complete these because of other priorities that arose but we are discussing on trying to get to work on these beginning of this year

How would you like to go about this? I think a lot of users have commented on the negative value to aggregate different metrics as one.

We have two main tickets that relate to what we plan to work on this:

• [Feat]: ability to "expand by" or similar for charts (e.g. expand by container, mount point etc) netdata-cloud#894 --> allow the expanding of charts based on some attribute, for example it could be the instance
• [Feat]: ability to save/override chart defaults in a room netdata-cloud#789 --> allow the storing of chart settings

That sounds good. Perhaps something like a filter on the right side could filter all relevant charts. For example if i put cgroup: somegroup, it would filter all. cgroups to only include that one.

Still, i prefer having a tree view on the right side ao i can easilly select all objects individually.

We are also working on improving the Custom Dashboards that will allow to make your own dashboards in an easier way

This is good. But needs more effort from users.

We weren't able to complete these because of other priorities that arose but we are discussing on trying to get to work on these beginning of this year

Perhaps you can communicate a roadmap on the features you indend to implement, I think this will ease users worry.

I hope you see how the current layout is not as usable as it was, otherwise it will be hard to take in our (users) needs.

That sounds good. Perhaps something like a filter on the right side could filter all relevant charts. For example if i put cgroup: somegroup, it would filter all. cgroups to only include that one.
Still, i prefer having a tree view on the right side ao i can easilly select all objects individually.

that is an interesting suggestion. we had discussed of two things that relate to this:

• extending the Global Nodes filter to allow filter also with chart labels and could perhaps be extended to the instances, this would filter across all the charts under Metrics tab that had that attribute
• improve the current TOC search since it is still quite basic

we'll see this probably after what I mentioned above

Perhaps you can communicate a roadmap on the features you indend to implement, I think this will ease users worry.

yes, we've been speaking internally of having some sort of public roadmap. let's see if we come around on releasing one

I hope you see how the current layout is not as usable as it was, otherwise it will be hard to take in our (users) needs.

I believe this is debatable, we have also gotten very good feedback of all the efforts done around the new charts, the NIDL framework, etc.
we understand some use-cases were impacted by this, hence why we are discussing about finding away to improve them, but there is still a lot of value on what we delivered

@hugovalente-pm

believe this is debatable, we have also gotten very good feedback of all the efforts done around the new charts, the NIDL framework, etc.
we understand some use-cases were impacted by this, hence why we are discussing about finding away to improve them, but there is still a lot of value on what we delivered

Fair enough. The Netdata chart system and the overall functionality is really great, especially the very powerful collector and retention possibilities. I was speaking from my own use case as a home user, where I wanted to quickly be able to check the status of my server and vms without having to drill down to see details. Because I know my systems pretty well, and being able to see individual services as separate charts, allows me to quickly see if trends are going the wrong way. This worked very well up until the new UI came along, which was quite disappointing.

That said, I did not mean to sound so harsh before. I'd be happy to help work out some more detailed work flows that would suit my use cases. Perhaps that should be a discussion for Discord or IRC?

🙏

I'd be happy to help work out some more detailed work flows that would suit my use cases. Perhaps that should be a discussion for Discord or IRC?

for sure, once we have done some initial brainstorm/prototyping from our side we'll let you know and get some feedback.
thanks for the honest share and being available to helps us out figuring the best way on this case

Opened Netdata this morning and got a message that I can only assume was intended to scare me into upgrading to a Enterprise subscription, insinuating that without it my machines were insecure.

And with that, I give up and just reverted to the old version of Netdata Debian ships as part of their repos.

Opened Netdata this morning and got a message that I can only assume was intended to scare me into upgrading to a Enterprise subscription, insinuating that without it my machines were insecure.

And with that, I give up and just reverted to the old version of Netdata Debian ships as part of their repos.

thanks for the feedback @foxt, the goal isn't to scare but to let people know how the Netdata Agent works by default. we provide information on our documentation for how you can secure your nodes by: Expose Netdata only in a private LAN, Use an authenticating web server in proxy mode, Use Netdata parents as Web Application Firewalls, etc.

all of the above require you to do these setups yourself, the Netdata Enterprise Agent is an offering we are exploring and see if people would be interested in such a thing were we would take some burden out of your hands

the goal isn't to scare but to let people know how the Netdata Agent works by default. we provide information on our documentation for how you can

The CTA wasn't 'read the docs', it was 'upgrade to Enterprise'. The link was not to the docs for what I have, it was to your pricing page.

The CTA wasn't 'read the docs', it was 'upgrade to Enterprise'. The link was not to the docs for what I have, it was to your pricing page.

Yes, we know the link wasn't to the docs. We are trying to measure the interest of such an offering that would replace the work a user would need to do by himself.
The docs are there for many time back, users interesting on securing their agents will probably have gone through those documents.

The docs are there for many time back, users interesting on securing their agents will probably have gone through those documents.

And those not interested, or those who have already read the docs, don't need a nagbar telling them to upgrade to Enterprise

I'd be happy to help work out some more detailed work flows that would suit my use cases. Perhaps that should be a discussion for Discord or IRC?

@Forza-tng are you still up on this offer? we would like to better discuss your use-case and expectations?

Well, I ended up removing Netdata from my servers.

The new UI isn't only overcomplicated, but also was causing high iowait.

As I basically was using the app as a web htop, I'll replace it with Glances via homepage instead.

Bye

@hugovalente-pm

I'd be happy to help work out some more detailed work flows that would suit my use cases. Perhaps that should be a discussion for Discord or IRC?

@Forza-tng are you still up on this offer? we would like to better discuss your use-case and expectations?

Hi, sorry for the late reply. Yes I would be interested in this :)

Well, I ended up removing Netdata from my servers.

The new UI isn't only overcomplicated, but also was causing high iowait.

As I basically was using the app as a web htop, I'll replace it with Glances via homepage instead.

Bye

I really don't get it, even if the new UI is overcomplicated (I would say enriched with tons of functionality, but as user myself when things change, it can bring me some inconvenience) we didn't stop shipping the old UI. Access it via your localhost:19999/v1.

Do you want it as default? overwrite the /usr/share/netdata/web/index.html with /usr/share/netdata/web/v1/index.html *

BEST solution; add a reverse proxy to point to v1 dashboard. Guys make it work for your case this is an open-source project and needs to evolve for all the users.

* you need to do that in every update

we didn't stop shipping the old UI

But you stopped maintaining it.

Do you want it as default? overwrite the /usr/share/netdata/web/index.html with /usr/share/netdata/web/v1/index.html *
* you need to do that in every update

And, see, you won't even natively allow making it default either.

Yes and this is expected, v2 is under development. v0, v1 are there, and we are open to community support.

And, see, you won't even natively allow making it default either.

I will commend on this note TTBOMK:
Adding an option wasn't a priority primarily because v1 is on community support.

Look I am not trying to convince you that our approach is right or wrong. But we are trying to keep this project going and evolving and we need to make certain decisions, some well thought others as logical side-effects. If you have any suggestions pitch your idea on a GitHub discussion https://github.com/netdata/netdata/discussions. Traction and healthy friction bring collaborative progress and change, not a goodbye. Always kindly saying :)

@tkatsoulas

Guys make it work for your case this is an open-source project and needs to evolve for all the users.

The new v2 UI has a proprietary license, it's not open source.

Please put it under an open source license if you claim to be an open source project or stop claiming netdata is an open source project if the only maintained UI is propietary!

I have the impression that all new developments are slowly rendering the v2 dashboard practically unusable without cloud integration.

Unless there is an option to configure locally (per instance) and persistently (not just during the browser session) for every chart, it will be painful to use.

For instance, defaulting to averaging all fan speeds or ping times is simply nonsensical, and reverting to defaults on every page reload is even more so - yet, that's how it is, with no signs of changing anytime soon. I do not want to choose relevant metrics every time again and again when I load the dashboard - I want it to be kept as I set it last time - without cloud and even when I open the dashboard in different browser.

In v1 (pre-v2) most metrics are displayed in more convenient way, there is rarely need to change something, but in v2 this is different - it is like to install and tune new OS every time when you open your laptop.

What is worse, all the changes introduced for v2, while they have their advantages, are making v1 and v0 less usable with each new release.

Even simple request - an option to show v1 by default, barely an hour of work - completely ignored or has so "low priority" that we will never see it.

All of this really feels like slowly boiling a frog - at some point, it will be so inconvenient to use Netdata in its default setup without cloud integration that every user will be forced to either use the cloud or choose another solution.

Most ironic, though, is the use of the killer feature - functions. It feels so strange that you have to subscribe to the cloud (free or not - it doesn't matter, it's still the cloud) for "privacy reasons." What? Do I have to register with a third party somewhere and share some private data (at least my IP address) just to use something that is generated locally? How is this "privacy"?

I have no problem when a project is leaning in the direction of "we need to make [more] money," but I have an issue when this is done covertly and non-transparently. Unfortunately, that's how it feels now, and I have no doubts that I am not alone in this sentiment.

Yes, I know, there is a free tier (and probably always will be) - but this is not an option because it is cloud. However small the amount (and nature) of data is sent somewhere - there are no guarantees that it will be properly protected and will not be used for anything else. Besides, it won't work at all in closed networks - and I am not talking about commercial usage, even home lab may be totally airgapped (and there is no free Netdata On-Prem).

Guys thank you so much for taking the time to express your thoughts/concerns/opinions, rest assured, I'll take them down after careful consideration and transfer them to the rest of the company (who are not following this thread).