TrueNAS SCALE Dashboard discussion.

[Damianos]

Greetings TrueNAS SCALE users!
One of the areas we're looking at improving in a future SCALE release is the Dashboard Page. We'd love to hear your ideas so please reply to this thread and answer the following questions.

1. Which widgets do you use the most?
2. Which widgets do you use the least?
3. How would you like to see the Dashboard page improved in TrueNAS SCALE?

 

[determined_hawking]

I'm going to skip the numbers because there are really two very distinct things I would like to see:

1) Better use of horizontal space when doing things like adding additional storage volumes to k8s Apps. It's really easy to get lost in the config with all of the vertical scrolling.

2) As alternative to #1, maybe the vertical would be fine if sections have better visual separation via alternating background colors or something.

3) Generally better information about the k3s Apps. ex: What's going on during a deploy? ex: If something fails, can I drill down into the Pod and Deployment details? If that stuff is there I've not found it and I end up SSH'ing in use kubectl.

EDIT: I'm generally a little bit surprised by the lack of action on this thread. But, hey, if I'm the only the poster that means my requests **definitely** get implemented, yes? ;)

 

[LarsR]

For 3.:
It would be nice if there was a gpu widget showing gpu stats from e.g. nvidia-smi, utilization of the gpu, clock speeds, if an app is using the gpu for transcoding etc. and for the Memory widget maybe a better differentiation between system processes and ram used by apps...

 

[Damianos]

@LarsR thanks for the GPU widget suggestion! You're not the first to request this.

@determined_hawking Thank you for your feedback as well. However this discussion is specifically for the front page dashboard. It sounds like your suggestions apply to the applications section of the UI. Please open a ticket on Jira and we'll gladly take it under consideration.

 

[determined_hawking]

@LarsR thanks for the GPU widget suggestion! You're not the first to request this.

@determined_hawking Thank you for your feedback as well. However this discussion is specifically for the front page dashboard. It sounds like your suggestions apply to the applications section of the UI. Please open a ticket on Jira and we'll gladly take it under consideration.

Oh, sorry. You are correct. I'm relatively new to the stack and didn't know about the Jira but I will check to see if there's anything over there and create a new FR if not. Thx.

 

[probain]

I made a post about where it could be improved. And these improvements would actually propagate throughout the entire system too, if done properly.
It also contains link to my suggestion Jira. Where I mention the same things.

Tldr; reduce the amount of padding. Condense the information to be just a tiny bit tighter, and things fit sooo much better.

Suggestion post (FIXED)

Update: Link is actual correct now
Update x2: Pinging @Damianos for visibility that the link is now correct

 

[chuck32]

You linked the wrong ticket I assume.

I was looking for your post, I remember reading it and thought about it when I saw this thread.

 

[probain]

You linked the wrong ticket I assume.

I was looking for your post, I remember reading it and thought about it when I saw this thread.

FFS. I should just stop trying to do any forum posting from my phone.

I'll fix the actual correct link immediately!
Thanks for letting me know

 

[Hutchca]

I'm not sure I actually "use" any of the widgets. I do glance at them frequently as I log in, before I jump to the page I need.
1. I tend to look at the Memory widget most often, also the Storage and Volume health indicators.
2. I never use The Help or System information, and rarely notice the Interface widget.
3. Maybe add widgets for Apps and Virtualization

 

[DanielYafacb]

Greetings TrueNAS SCALE users!
One of the areas we're looking at improving in a future SCALE release is the Dashboard Page. We'd love to hear your ideas so please reply to this thread and answer the following questions.

1. Which widgets do you use the most?
2. Which widgets do you use the least?
3. How would you like to see the Dashboard page improved in TrueNAS SCALE?

Hi
1. Memory
2. Interface
3. Last 10 LOGS messages from virtual machines

 

[Perry The Cynic]

I've pretty much given up on the dashboard because it doesn't tell me the things I really need to know (yay, Prometheus). But since you asked, in rough order of importance (to me):

• Generally, switch from "this moment" stats to "recent average plus peaks/saturations" views. It's pointless watching the CPU indicator jump from 15% to 95% and back (snazzy but distracting). Show me the 1-minute average and tell me if we've hit saturation (100% CPU load). The same applies to network and interface - don't (just) tell me the current throughput; tell me I'm averaging 27% of capacity and the 1-hour peak was 95%.
• Add an apps summary panel. What are my top RAM-consuming apps? What are my top CPU-consuming apps? Any uncommanded pod restarts? Flag any apps that are not in running or stopped state.
• Add a dataset panel, summarizing the most important datasets (almost full, highest activity, etc.)
• ARC hits and misses summarized. No fancy details (that belongs in another section that, uh, doesn't seem to exist.) But there's a world of difference between 99% ARC hits and 90%. Also, L2ARC summary. None of that is pool specific, so I guess this would go to memory (or a separate, new "zfs" panel).
• Break the "services" part of the memory panel into "system" and the top few apps (plus "other apps").
• Add an events summary panel. Even if the jobs/alerts displays were great (they aren't), I'd still like to see notable recent events (whether they're errors or not). And no, the three-log-lines-at-the-bottom thing isn't helping; I want notable events, not some implementation detail logs.
• Break the CPU numbers into components - how much goes to system, how much to apps, which are the top 1-2 apps.
• Lose the per-CPU load display. It's nigh pointless (unless you pin apps to CPUs, which I consider too specialized for the top level).
• The CPU temperature display should be anchored to acceptable high temps. As it stands, is 58° fine or almost too hot? Just use a car-style indicator with a number in it. Be opinionated on what's "normal".
• Transfer summary. How much data was replicated (or rsync'ed, or cloud-synced) in and out in the last day? What are the biggest contributors?

Cheers
-- perry

 

[kurzweilfreak]

Hi,
Comments from a newbie:
1. Memory widget, Storage and Volume health indicators
2. System and Help widgets
3. I liked the above comment about “average recent CPU (and GPU) graphs”, and also list of top cpu/memory/gpu-using apps/vm (maybe two separate widget - top 5 +”other” showing the average across the other apps)

Cheers

 

[unorthodox]

1. Network, then Memory.
2. Interface
3. 
   - An annoyance with the current dashboard is the Network widget text (below Link State Up), and the graph (on hover) units of measurement not matching. One is given in "KiB/s" and the other just "M".
   - Another gripe with the current dashboard is the Pool widget, when clicking on the "Data" field, it lists the drives vertically (one per line) when there seems to be enough space to list them as a comma separated list, and it wouldn't cause this broken looking layout with text protruding below the line when you have more than 3 drives in the pool.
   - In general I agree with probain that there is too much padding in the entire UI. Specifically for the dashboard it would be nice if the widgets would use the whole width of the dashboard and dynamically optimize horizontal space usage (I have a combined 3 inches of blank space left and right of the widgets). In the same vein, how often do people create new pools to require a big button for it in the Storage widget?
   - The CPU widget seems pointless unless you just sit there and watch your NAS do what it does.
   - I love Perry The Cynic's idea of showing ARC hits and misses this seems important for a ZFS NAS.
   - I could see the most recent event for each data protection task that ran (or tried running) being useful (with the ability to only show select tasks, or failed tasks), knowing how long each scrub, replication, etc. took to complete would also be nice.

 

[probain]

The memory circle-chart could have a fourth (or more) for other types as well, e.g. L2ARC, apps, VMs. Instead of just having three, and lots of them being clumped together under the term "services".

These could be on an inside diameter of the main three?

I also like the idea of panels for VMs and apps. However, this should show useful information more than just on/off status.

But the main one for usability, is the whole UI. That I believe could use some slight trimming. Just going from 100% layout zoom, to 90%. Makes the entire screen fit disproportionately more information. It's crazy. And this is something that, if done right. Would benefit the entire UI throughout TN.

 

[Perry The Cynic]

Yeah, breaking out the "zfs cache" memory slice into the parts dedicated to L2ARC and dedup tracking (and metadata) would be nice as well. Note how this chart does (or should do) double duty:
(1) How much of my RAM is used by zfs overall (vs. apps and system); and
(2) How much of my ARC is actually used to cache payload data (and metadata), vs. a price I pay for "advanced" zfs features.

Again, this is a top-level glance interface, so don't overload the memory display with details (make a separate zfs panel for that). But breaking the "zfs cache" into "real cache" and "overhead" would be quite glanceable. Note how due to ARC compression, the "currently used" ARC data portion is different from the consumed RAM, so this might not fit into one simple round thingie.

Cheers
-- perry

 

[nobillgates]

Greetings TrueNAS SCALE users!
One of the areas we're looking at improving in a future SCALE release is the Dashboard Page. We'd love to hear your ideas so please reply to this thread and answer the following questions.

1. Which widgets do you use the most?
2. Which widgets do you use the least?
3. How would you like to see the Dashboard page improved in TrueNAS SCALE?

1. Storage
2. Help
3. A Summary widget that has key metrics from Storage (e.g. % used), Network and Memory - or a choice of metrics from any main widget.

MTIA
;-}
P

 

[mada kamal]

Greetings TrueNAS SCALE users!
One of the areas we're looking at improving in a future SCALE release is the Dashboard Page. We'd love to hear your ideas so please reply to this thread and answer the following questions.

1. Which widgets do you use the most?
2. Which widgets do you use the least?
3. How would you like to see the Dashboard page improved in TrueNAS SCALE?

Hi there is your answers
1. Memory
2. Interface
3. Last 10 LOGS messages from virtual machines

 

[Damianos]

Thanks everybody for sharing their thoughts. It has been enlightening and we will take them into consideration for any upcoming dashboard improvements.

I have one final question about phones and smaller screens.

Do you find the mobile layout for the Dashboard page useful or would you rather it just be the regular widget cards in a single column?

 

[probain]

Thanks everybody for sharing their thoughts. It has been enlightening and we will take them into consideration for any upcoming dashboard improvements.

I have one final question about phones and smaller screens.

Do you find the mobile layout for the Dashboard page useful or would you rather it just be the regular widget cards in a single column?

Personally I find the native mobile UI pretty much useless. And therefore I've opted to use the desktop layout even when browsing mobile.
Basically the mobile UI is spilt up to much. Dashboard should be the overall at a glance pane. If I need to view something more specific, I'll click into the relevant section. But mobile right now pretty hides everything.

So I would much prefer a pane with widgets in a single column, being right at the start.

 

[Damianos]

Personally I find the native mobile UI pretty much useless. And therefore I've opted to use the desktop layout even when browsing mobile.
Basically the mobile UI is spilt up to much. Dashboard should be the overall at a glance pane. If I need to view something more specific, I'll click into the relevant section. But mobile right now pretty hides everything.

So I would much prefer a pane with widgets in a single column, being right at the start.

Thanks for sharing your feedback.