Show HN: Monitoror – Unified monitoring wallboard • Hacker News Board

Re: Show HN: Monitoror – Unified monitoring wallboard

#81

Post by achow » Wed, Mar 04, 2020, 5:15 PM UTC

wjossey wrote: post #24

CSDude wrote: post #6

I know people like wallboards and monitors but we found them anti-pattern. If you find yourself looking at a wallboard/dashboard, it should already be an automated alert.

Strongly disagree. Understanding your metrics is a key part of so many roles, from devops, to product teams, to marketers... Yes, you should be automating alerts whenever possible. Yes, you should be putting up key metrics in a visible place so everyone can see how the product is performing. I can’t tell you how many times I caught an issue because I knew our metrics backwards and forwards, but it didn’t trip an aler…

Strongly agree (with you).

From the very first formulation of Ubiquitous Computing, the idea of a calmer and more environmentally integrated way of displaying information has held intuitive appeal. Weiser called this “calm computing”.. When information can be conveyed via calm changes in the environment, users are more able to focus on their primary work tasks while staying aware of non-critical information that affects them. Research in this sub-domain goes by various names including “ambient displays”, “peripheral displays”, and “notification systems”...

A Taxonomy of Ambient Information Systems: Four Patterns of Design

https://www.cc.gatech.edu/~john.stasko/papers/avi06.pdf

Re: Show HN: Monitoror – Unified monitoring wallboard

#82

Post by C1sc0cat » Wed, Mar 04, 2020, 5:18 PM UTC

tyrust wrote: post #72

C1sc0cat wrote: post #55

Earlier quoted context omitted.

An automated email is ok but seeing visually a graph flat line or a monitor turn red is much more likely to get noticed.

If you ignore alerting then it's likely that your alerts are too noisy. See "alert fatigue".

Its that nice Mr Googles alerts

Re: Show HN: Monitoror – Unified monitoring wallboard

#83

Post by reaperducer » Wed, Mar 04, 2020, 5:20 PM UTC

SergeAx wrote: post #77

ampdepolymerase wrote: post #68

Earlier quoted context omitted.

But...but.. something.. something Stallman...vendor lock-in...closed-platforms bad...something.

If it is a sarcasm, then please mind that original comment author got really humped by this app's vendor when it stopped working. Maybe Stallman got something right after all?

It didn't stop working. It just doesn't get updated anymore.

Re: Show HN: Monitoror – Unified monitoring wallboard

#84

Post by OJFord » Wed, Mar 04, 2020, 5:26 PM UTC

CSDude wrote: post #14

OJFord wrote: post #12

Earlier quoted context omitted.

What would your alert be for # open PRs (an example in the demo linked from posted page)? How often would it fire? Whatever the answer, that's a different thing from this. Both have their place.

If you just want to have a nice visualization to look at some numbers, fine. But, if you want to detect problems, it's ineffective. I saw too many companies do it to actually monitor the state of things and find out problems with charts, numbers, traffic lights etc.

But that's my point, it isn't for alerting about problems, some things have a 'status' that might be interesting, but isn't a problem, or something to fire an alert on necessarily.

You could have unintrusive notifications (inaudible etc.) to 'alert' to such statuses I suppose, if they were kept in view and not 'dismissed' (whatever that means for the medium they came in) - but then really you're just implementing a version of something like this Monitoror in your inbox, phone notification tray, Telegram channel, or whatever.

You're not going to rip out logging, prometheus, or services' that this connects to own UI just because you have alerting, so I don't see why you would this. It's like prometheus & grafana for higher level stuff. (Of course you could use those tools for this sort of monitoring too, but that's not really the point.)

Re: Show HN: Monitoror – Unified monitoring wallboard

#85

Post by djsumdog » Wed, Mar 04, 2020, 5:47 PM UTC

stiray wrote: post #80

SergeAx wrote: post #50

Earlier quoted context omitted.

But why do you need an app where webpage is more than enough?

One more thing, if I want a webpage, I will just write plugin for this: https://github.com/netdata/netdata It is meant as system monitor but it can chew up preatty much anything. And it is fast and lean. Really fast.

Have you found any tools for writing custom dashboard in netdata?

The docs are really light and custom dashboards seems to involve hacking the default HTML and pulling out all the components (javascript/divs) you need. I thought about writing my own, but have too many projects already.

Re: Show HN: Monitoror – Unified monitoring wallboard

#86

Post by RBerenguel » Wed, Mar 04, 2020, 5:51 PM UTC

CSDude wrote: post #6

I know people like wallboards and monitors but we found them anti-pattern. If you find yourself looking at a wallboard/dashboard, it should already be an automated alert.

I've spotted interesting "things" from idly looking at our dashboard while chatting with coworkers (and more than a few were interesting enough to warrant a lot of investigation and double-checking of metrics, providers and stack). They were not alert-able, or not very easily unless we wrote some complex time series analysis system for our internal metrics.

Re: Show HN: Monitoror – Unified monitoring wallboard

#87

Post by RBerenguel » Wed, Mar 04, 2020, 5:55 PM UTC

TeMPOraL wrote: post #63

wpietri wrote: post #42

Earlier quoted context omitted.

Not at all. Alerts serve a different purpose. One of the most important things a team needs over the long haul is a feel for their system. Many people refer to this as mechanical sympathy. And the way you develop that is long-term exposure to rich data. Alerts are the red and yellow lights on your dashboard. But you get mechanical sympathy by listening to the sound of the engine, feel of the road, and the smell of th…

> you have to have the right information shown in the right ways for people to develop a correlative, intuitive understanding of what they've built Lots of things have to be right for this to work, unfortunately, and company dashboards I've seen so far tend to be nowhere near it. For instance, the dashboard refreshed $PERIOD only makes sense if you're showing data that updates $PERIOD, and if you can respond to chang…

Likewise! I found this some time ago: https://dl.acm.org/doi/10.5555/1045502.1045526

but never tried it, I just left it as a fun idea for the future.

Re: Show HN: Monitoror – Unified monitoring wallboard

#88

Post by jedberg » Wed, Mar 04, 2020, 6:09 PM UTC

kube-system wrote: post #67

jedberg wrote: post #51

Earlier quoted context omitted.

That’s... actually true. It doesn’t matter how fast you’re going unless you’re speeding. And it distracts you by making you look down. The only reason we don’t have that yet is because the car doesn’t know the speed limit everywhere all the time.

I strongly disagree. I can think of a ton of reasons why a driver may need (or even be legally required) to know their speed regardless of speed limit: * when speed restricted by equipment (trailer, temporary spare, etc) * when observing advisory speeds * when observing minimum speed requirements * as a reference for judging appropriate speeds under inclement conditions * as a reference for judging appropriate accele…

Of course an alert system would have to be able to understand all those things. That's why we don't have that kind of system.

A single number in isolation is rarely useful. Graphs with trends are useful. Alerts are useful.

The only reason we don't have alert based speeds is because it can't get all the necessary information to make a useful alert, so we compromise by telling you the number.

> as a reference for judging appropriate acceleration/deceleration rates when entering/exiting the roadway

A perfect example of why a graph would be ideal here, not a single number.

Re: Show HN: Monitoror – Unified monitoring wallboard

#89

Post by stiray » Wed, Mar 04, 2020, 6:10 PM UTC

djsumdog wrote: post #85

stiray wrote: post #80

Earlier quoted context omitted.

One more thing, if I want a webpage, I will just write plugin for this: https://github.com/netdata/netdata It is meant as system monitor but it can chew up preatty much anything. And it is fast and lean. Really fast.

Have you found any tools for writing custom dashboard in netdata? The docs are really light and custom dashboards seems to involve hacking the default HTML and pulling out all the components (javascript/divs) you need. I thought about writing my own, but have too many projects already.

Set up netdata, check which plugin serves the type of chart you want, find shortest of that type, make a copy and hack it untill it works. I was doing it a while back so I cant remember the details but it wasnt something special to do... Ignore the docs, existing plugins are all the documentation you need.

Re: Show HN: Monitoror – Unified monitoring wallboard

#90

Post by zola » Wed, Mar 04, 2020, 6:21 PM UTC

jaywalk wrote: post #71

jedberg wrote: post #51

Earlier quoted context omitted.

That’s... actually true. It doesn’t matter how fast you’re going unless you’re speeding. And it distracts you by making you look down. The only reason we don’t have that yet is because the car doesn’t know the speed limit everywhere all the time.

I have that in my 2020 Ford, and you can tune the alert threshold from 0-5mph over the limit. I wouldn't even consider turning it on unless I could set it to at least 10mph, however. The vehicle has a camera that looks out for speed limit signs, and then updates a little icon on the instrument cluster with the current speed limit. It works very well.

I was driving a car like that few times, with beeping when the limit was exceeded. Annoying, but informative, unless road sign was limiting vehicle mass to 10t and not speed to 10km/h. It was carsharing vehicle, so I didn't bother to turn it off.