DRUIDS: Datadog Reusable User Interface Design System(druids.datadoghq.com)
druids.datadoghq.com
DRUIDS: Datadog Reusable User Interface Design System
https://druids.datadoghq.com/?ref=hackernews
106 comments
Awwwwwwwwwwww... that makes me SO sad. I LOVE Datadog's dashboards and UI and was so excited that they opened it up. Finally some competition for MUI, I thought, but nope :(
> Finally some competition for MUI, I thought
NPM is full of every other company's internal component library they open source/release publicly. Datadog would have been nothing special.
NPM is full of every other company's internal component library they open source/release publicly. Datadog would have been nothing special.
Yeah? What are some good ones?
from the top of my head...
* Sales Force - Lightning
* Atlassian - AtlasKit
* Microsoft - Fluent UI
* Uber - BaseWeb
* SAP - UI5 Web Components for React
* Palanti - Blueprint
* AWS - Cloudscape
* Sales Force - Lightning
* Atlassian - AtlasKit
* Microsoft - Fluent UI
* Uber - BaseWeb
* SAP - UI5 Web Components for React
* Palanti - Blueprint
* AWS - Cloudscape
Cool, thanks, I'll have to check these out!
I don't know why they couldn't put a friendly version of that message on the landing page.
Are people outside the org even meant to see the page? That is odd.
Well there is a ref=hackernews in the URL for tracking so I guess yes ?
Key factor here that people should know.
I tried out this design system but got a bill for $125,000 for scrolling charges
Context?
It's really easy to run up a huge bill from Datadog if, for example, some engineer adds a whole lot of new tags/dimensions to metrics data because they think it would be useful to have in the future (full disclosure: that was me, in a previous role. I think I "spent" almost $30k on extra metrics before anyone realised and we tracked down what happened - I was new to DD and didn't even know they charged extra for those things)
What was the size of this data set? Curious as to just how much data you have to handle to incur a bill like that.
They charge based on the number of unique combinations of metrics and tags. So if you were to add a tag for something with high cardinality, like say the source ip address, or even a browser user agent, you can ring up a pretty high bill without even sending that much data, because you have a large number of metrics.
This is pretty much what happened, but in my case it was a bunch of IDs for objects owned by customers. It was intended (and did) to provide detailed performance metrics related to our Elasticsearch searches, but in the end I had to drop a lot of them due to cost.
Really awesome work on the "examples" section. As someone personally working on a design system docs, I admire the team's thoughtfulness on building that
WOW.
This is some of the best documentation I have ever seen, and a very elegant design, too.
WOW.
I am working on documentation for my own product right now, and this is inspiring.
This is some of the best documentation I have ever seen, and a very elegant design, too.
WOW.
I am working on documentation for my own product right now, and this is inspiring.
Other great sources of DS docs:
https://atlassian.design
https://carbondesignsystem.com/
https://polaris.shopify.com/
Obv for non-DS docs, Stripe's docs are gold medalists.
https://atlassian.design
https://carbondesignsystem.com/
https://polaris.shopify.com/
Obv for non-DS docs, Stripe's docs are gold medalists.
Atlassian is only nice for what they actually display on that page, but their design system (the components at least) is actually 2x more expansive, meaning you are out of luck if you need info on any of that.
Of course, they do give you the source, so that’s nice.
Of course, they do give you the source, so that’s nice.
I guess mobile support isn’t part of those design systems. The Atlassian docs don’t work on mobile at all. The other two are at least functional, but look really bad.
This is the difference when you have dedicated, well-funded teams for which the design is the product. For most developers, it's something they are made to do on the side - I mean, "but you're a full stack developer, right??".
OK, I am not one for complaining about designs, and I'm not even going to say this is bad.
But for some reason, I can't even look at the page. It's giving me a headache, just a few seconds of looking at it makes me feel... very off. Almost feel like I'm staring at an optical illusion. Super weird.
But for some reason, I can't even look at the page. It's giving me a headache, just a few seconds of looking at it makes me feel... very off. Almost feel like I'm staring at an optical illusion. Super weird.
I can really appreciate the weirdness of the aesthetic in an age where every startup has to make everything cute and fashionable
Besides the product itself, I can appreciate the lengths gone to make DRUIDS work as an acronym, and the more fun UI/logo elements present.
I'm curious, what is it that drives every tech company to eventually publish a UI framework? I get the value of having an internal UI framework which allows anyone in an organisation to quickly throw something together which is at least vaguely in line with branding and UI patterns, but what value do they get from then making that available to the general public. Surely it just puts a burden on the maintainers because they can no longer just send a quick email or Slack message to the relevant channel saying "we're going to break backwards compatibility for widget X, make sure you update".
> I'm curious, what is it that drives every tech company to eventually publish a UI framework? I get the value of having an internal UI framework which allows anyone in an organisation to quickly throw something together which is at least vaguely in line with branding and UI patterns, but what value do they get from then making that available to the general public.
It makes their front-end engineers and designers happy and acts as a recruiting tool: Look at what we're building for internal use and our culture of open source contributions.
> Surely it just puts a burden on the maintainers because they can no longer just send a quick email or Slack message to the relevant channel saying "we're going to break backwards compatibility for widget X, make sure you update".
It sometimes does, if they bother to support public issues vs. it being available but mostly only supporting internal use cases.
It makes their front-end engineers and designers happy and acts as a recruiting tool: Look at what we're building for internal use and our culture of open source contributions.
> Surely it just puts a burden on the maintainers because they can no longer just send a quick email or Slack message to the relevant channel saying "we're going to break backwards compatibility for widget X, make sure you update".
It sometimes does, if they bother to support public issues vs. it being available but mostly only supporting internal use cases.
One place I worked did that because it made development/packaging easier. You didn't need to maintain a private repo with auth for something that eventually gets published publicly anyway
> I'm curious, what is it that drives every tech company to eventually publish a UI framework?
1. Low risk. Not likely to create competitors or give away the secret sauce.
2. Allows partners to build lookalikes/similar look and feel apps.
3. Enables vendors/partners to contribute in a meaningful way that helps both parties.
1. Low risk. Not likely to create competitors or give away the secret sauce.
2. Allows partners to build lookalikes/similar look and feel apps.
3. Enables vendors/partners to contribute in a meaningful way that helps both parties.
I always wonder why you need a design system for a dev first product. MaterialUI + theme palette would be enough most probably but I guess its the (soon to be over) age of free money.
I hear you on "don't build your own design system"... but Datadog's got a really complex UI, I think they've definitely graduated past the point where something like MUI would be a good choice (and obviously they have the resources to do an internal design system, and do it well).
For Datadog I think it makes sense.
For Datadog I think it makes sense.
Material UI... in 2022? Why not just return to Bootstrap?
what's wrong with material UI? it's fine for people building admin panels, ERP and what not AKA real apps, not websites. What would you use instead?
Literally anything. Unstyled HTML tags are better
Using something like Material UI is always a great idea to get you moving quickly. But eventually you reach an annoyance point once you scale up the number of components you're creating. Especially if you need more customized things. The layers of hacks and work around start to get confusing. And new folks give you lots of WTF comments.
Same reason people build new libraries or languages. They feel limited by the "thing" they're using. :)
Same reason people build new libraries or languages. They feel limited by the "thing" they're using. :)
I cannot wait for material ui to die and go by the wayside
It's fine for what it is - a batteries included component library. Not a lot of the current ones come close. Maybe Mantine or Semantic but there's not a lot of great ones out there.
What is better if you want to build admin GUI and you want to spend little time with style sheets?
I've used both and don't find Antd to be better than mui. I find mui more enjoyable and more feature full but it seems to be a marginal difference.
Use tailwind with the most basic class applications.
I was JUST thinking of building a small web app for viewing local structured JSON logs with a subset of the features in the Datadog logs explorer. It'd be a nice little bonus to build the UI with the same components!
I find datadog's UX to be one of the least user friendly and find myself constantly frustrated by how inconsistent it is. Looking at a time span in logs? Click a link to a trace in APM and it's reset.
Looking at process cpu use and there's no time scale...
Some visualizations you can cross filter on, some you can't. Some only hide things in that specific visualization.
It's also very slow.
It also doesn't work at all on mobile.
Not sure why anyone would be jumping to use this.
Looking at process cpu use and there's no time scale...
Some visualizations you can cross filter on, some you can't. Some only hide things in that specific visualization.
It's also very slow.
It also doesn't work at all on mobile.
Not sure why anyone would be jumping to use this.
Isn't DRUIDS a reference to something? I swear I remember it being from a TV show, but I can't quite remember.
Druid is a religious term but nowadays most commonly used in fantasy fiction, don’t know it’s first written usage. It is a term that has been around since ancient Roman times. Particularly England with the Wicca religions.
Imagine a forest dwelling witch (in very basic terms).
Imagine a forest dwelling witch (in very basic terms).
Say what you want about Datadog's pricing and sales tactics... the product is a joy to use.
When my team was forced to use it a few years ago it was order of magnitude more expensive than diy prometheus/grafana while being less friendly to devs - their metric query language absolutely sucked. Was more friendly to non-devs who liked pretty ui tho…
We also had some collector troubles and support basically did nothing but wasted our time in calls repeatedly
We also had some collector troubles and support basically did nothing but wasted our time in calls repeatedly
Managed services are always more expensive than DIY FWIW. You're paying to make running the product someone else's problem.
That is why DIY is usually more expensive than managed services. Engineering hours are expensive and best spent on your core competencies.
DIY only make sense at a very small scale or very large scale, everything in between is usually best offloaded to those which do it as their core competency.
DIY only make sense at a very small scale or very large scale, everything in between is usually best offloaded to those which do it as their core competency.
I would caution against sweeping generalizations like that. In this case “diy” part is basically just configuration management which with dd you will have to do anyway. And sure they make it slightly easier by providing defaults for most things but Prometheus/grafana do a decent job at it too.
More broadly I’ve never used managed service that would “just work” and wouldn’t require substantial configuration and often times bunch of workarounds but maybe those exist
More broadly I’ve never used managed service that would “just work” and wouldn’t require substantial configuration and often times bunch of workarounds but maybe those exist
At every single org I've been where Datadog has been considered, the conclusion has been "Yes, it would be cool, but we really can't justify the price."
Yes, in theory, in the middle scale, you should outsource things, but in practice, it only works if the managed service is at the right price.
Yes, in theory, in the middle scale, you should outsource things, but in practice, it only works if the managed service is at the right price.
Why does DIY make sense at very small scale IYO?
It seems like very small scale has the highest leverage of utility-priced services (and often fits into free tiers of many).
It seems like very small scale has the highest leverage of utility-priced services (and often fits into free tiers of many).
We were not at a scale where you would want to hire dedicated observability expert but with their pricing it totally made sense. My guess is their play is to get in early and get you locked in gud
They have a good product, but no matter how good it is, after the experience I had with their sales, I will never use the service again.
their sales is abysmal. i have a new account manager every 6 months that wants to schedule a meeting. they put stuff on proposed contracts where they don’t even offer a discount.
just terrible.
imo they should drastically simplify their billing dimensions so a simple human can understand it. for a certain size of company it just makes no sense to need to be engaged with a sales teams.
just terrible.
imo they should drastically simplify their billing dimensions so a simple human can understand it. for a certain size of company it just makes no sense to need to be engaged with a sales teams.
They're overly aggressive and the cold calls to my personal cell mean I'll avoid their product whenever I'm in a decision making role.
would you mind elaborating a little bit on what happened?
Signed up for their service to get visibility into our infrastructure; we're a small company with a big setup. They bill hourly, but we do a lot of small instances that run for a few minutes at a time. Twelve instances running 3 minutes each is billed as 12 hours of monitoring.
We approached them to see if they would work with us on reducing the massive bill that resulted. They agreed to cut it by 50% if we signed up for additional services. I'm not talking about a future volume discount; we were working with them for a good faith credit once we discovered the mismatch with their billing model (we had already filtered out those instance types)
Objectively, we owed the money. However, every other vendor I've run into works with small companies like ours without resorting to those kinds of tactics, so it's a pretty terrible look for them.
We approached them to see if they would work with us on reducing the massive bill that resulted. They agreed to cut it by 50% if we signed up for additional services. I'm not talking about a future volume discount; we were working with them for a good faith credit once we discovered the mismatch with their billing model (we had already filtered out those instance types)
Objectively, we owed the money. However, every other vendor I've run into works with small companies like ours without resorting to those kinds of tactics, so it's a pretty terrible look for them.
Not only do they cold call people's personal phones, they do it after being told no. In all of their communication, they are pushy and give off used car salesman vibes.
I have no experience with sales but I always wonder what kind of incentives these people are given to take such draconian measures. Are they acting like stalkers because they get a fat commission check or is it typically do to something else?
They need to meet a quota. But also commission paid.
When it works, it is awesome.
But there are some caveats. Facets can break in unexpected ways and the last time you want to be dealing with this is when you're dealing with a fire in production.
But there are some caveats. Facets can break in unexpected ways and the last time you want to be dealing with this is when you're dealing with a fire in production.
Agreed. I had the same experience though as many others when it comes to Sales. I understand it is a complex product but they couldn't demo me anything even after 2 meetings. They wanted a 3rd meeting for the demo even though I made it clear on the 1st meeting that I am only interested in specific products (log monitoring etc) and would be good to see a demo in 2nd meeting.
Too much friction in their sales process. But I guess I am not the target audience.
Too much friction in their sales process. But I guess I am not the target audience.
Re: not target audience — I think you're right, I am also part of that cohort (speaking as an engineer at least).
This is one of the reasons why I steer away from anything that requires a demo. If an org can't present even a read-only interactive version of the product, then it likely means that there is a KPI/OKR-heavy pitch intended for management or non-engineering business stakeholders to hear (of which the upselling you alluded to is a part).
The majority of the (F)OSS alternatives out there can be demo'ed with little-to-no engineering lift from prospective users. This is meaningful for the adoption story because it creates bottom-up pressure to internally pitch to relevant stakeholders— a much more powerful tool than external pitches. The fact that Datadog seems either unwilling or incapable of doing this historically, while touting one of the more expensive products in that particular vertical, suggests that the product value-add may not speak for itself (at least to a significant subset of engineers).
This is one of the reasons why I steer away from anything that requires a demo. If an org can't present even a read-only interactive version of the product, then it likely means that there is a KPI/OKR-heavy pitch intended for management or non-engineering business stakeholders to hear (of which the upselling you alluded to is a part).
The majority of the (F)OSS alternatives out there can be demo'ed with little-to-no engineering lift from prospective users. This is meaningful for the adoption story because it creates bottom-up pressure to internally pitch to relevant stakeholders— a much more powerful tool than external pitches. The fact that Datadog seems either unwilling or incapable of doing this historically, while touting one of the more expensive products in that particular vertical, suggests that the product value-add may not speak for itself (at least to a significant subset of engineers).
Are you kidding? It's visual vomit and takes 3-4 clicks to get to relevant data. The only "great" thing about it could be the tracing but something you can easily get with OpenTracing/Jaeger. I have to use Datadog daily and sorely miss Grafana.
I wasted my time reading the release post.
Datadog's product is a bit too close to Apache Druid to have named their design system so similarly.
From https://druid.apache.org/ :
> Druid unlocks new types of queries and workflows for clickstream, APM, supply chain, network telemetry, digital marketing, risk/fraud, and many other types of data. Druid is purpose built for rapid, ad-hoc queries on both real-time and historical data.
From https://druid.apache.org/ :
> Druid unlocks new types of queries and workflows for clickstream, APM, supply chain, network telemetry, digital marketing, risk/fraud, and many other types of data. Druid is purpose built for rapid, ad-hoc queries on both real-time and historical data.
I'm really curious what kind of UX riddle this "Mystery of the DRUIDS" everyone keeps talking about is.
Does anyone know/recognize what this documentation site is built with? It's fantastic.
Custom, perhaps?
Custom, perhaps?
It definitely looks custom.
Is this the same Datadog that sends spam calls to developers after office hours to sell their tools?
Yup. I find their sales strategy deplorable as not only do they cold call like crazy, but their presence at conferences are all sales and no meat.
For example early on in AWS Lambda’s life, DataDog was hosting a session at reInvent that looked like a semi-advanced dive into the new technology. Awesome! I was legitimately excited and thought this might be one of the better sessions of the conference. I show up only to find it is 30 minutes of stand up comedy, 10 minutes of the most basic “how to create a lambda function” tutorial (probably ripped right from Jeff Barr’s blog), and 15 minutes of “you should buy DataDog”.
To this day, we use “DataDog” as in team meetings as a term to communicate shadiness etc.
(Edited to fix typo on Barr’s name)
For example early on in AWS Lambda’s life, DataDog was hosting a session at reInvent that looked like a semi-advanced dive into the new technology. Awesome! I was legitimately excited and thought this might be one of the better sessions of the conference. I show up only to find it is 30 minutes of stand up comedy, 10 minutes of the most basic “how to create a lambda function” tutorial (probably ripped right from Jeff Barr’s blog), and 15 minutes of “you should buy DataDog”.
To this day, we use “DataDog” as in team meetings as a term to communicate shadiness etc.
(Edited to fix typo on Barr’s name)
Their billing practices aren’t great either. Non transparent pricing, requiring docusign after signup to change plans, and no refunds for unused services.
Not to mention that, by default, they bill by the whole calendar month for each infrastructure and APM host. Scale your Kubernetes cluster up and then back down? That'll be an extra $18 + $36 per additional node (not $15 + $31 – that's the contract pricing, not the on-demand pricing), even if they were only online for a few days – even if they were only online for thirty seconds. Swap out a node? By default they bill by unique instances, not by number of instances, so they'll bill you for that, too.
If you ask them about it, they'll “happily” put you onto hourly on-demand billing (which seems to fix the unique vs. count thing, too), which is more expensive if you let something run on-demand for a whole month... but isn't the point of an on-demand service that you're not running it for a whole billing period?
Also, their agent logs fairly noisily, and of course its logs count toward your quota! I upgraded a cluster without also upgrading the agent, and didn't notice for about a week that each agent was happily spamming away about some long-deprecated Kubernetes API no longer being available[0]. At $2.55/million log lines and fewer than a million lines logged, this was not a costly mistake, but it's the principle of the thing. Why should an incompatibility in their agent (which their dashboard could specifically alert about, but doesn't!) cost me money?
[0] https://github.com/DataDog/helm-charts/issues/620#issuecomme...
If you ask them about it, they'll “happily” put you onto hourly on-demand billing (which seems to fix the unique vs. count thing, too), which is more expensive if you let something run on-demand for a whole month... but isn't the point of an on-demand service that you're not running it for a whole billing period?
Also, their agent logs fairly noisily, and of course its logs count toward your quota! I upgraded a cluster without also upgrading the agent, and didn't notice for about a week that each agent was happily spamming away about some long-deprecated Kubernetes API no longer being available[0]. At $2.55/million log lines and fewer than a million lines logged, this was not a costly mistake, but it's the principle of the thing. Why should an incompatibility in their agent (which their dashboard could specifically alert about, but doesn't!) cost me money?
[0] https://github.com/DataDog/helm-charts/issues/620#issuecomme...
Their billing doesn't match up with AWS's (AWS is by the second, Datadog by the hour, or at least when we used it) even though it's by the cloud instance (doesn't roll over). So we ended up paying more for the monitoring than the actual resources being monitored. When we asked for a break, they agreed to give us a 50% break IF we signed up for additional services.
Yes, their sales team is far too aggressive, I can't imagine it's successful for building their brand with developers
I too have had more spam calls from Datadog than any other tech company. Their product seems great but after what feels like harassment, I’ve never wanted to give them my money.
As someone allergic to those stupid cold calls, this is real disappointing to hear—I hadn't been exposed, presumably because I'm already a customer and have been for years.
As much as I love the product, I'll have to reconsider my usage of Datadog in future projects.
As much as I love the product, I'll have to reconsider my usage of Datadog in future projects.
I'll never use Datadog for this reason. I have been pestered by so many salesmen _relentlessly_, even after saying I was not interested.
How does one go about removing a phone number off of these sales data aggregators?
I don't think I've ever explicitly given these phone numbers to tools like this (e.g. signing up to Datadog with a phone number), so this seems like sensitive PII that must have been leaked and scraped in some shady source that these sales "data-enrichment" tools happily take.
I don't think I've ever explicitly given these phone numbers to tools like this (e.g. signing up to Datadog with a phone number), so this seems like sensitive PII that must have been leaked and scraped in some shady source that these sales "data-enrichment" tools happily take.
There are probably way too many data aggregators out there to keep track of completely, but theres definitely a few github repos I've seen that keep lists of both the companies and their opt-out procedures (some with automation).
This is one of the better ones I've seen: https://github.com/yaelwrites/Big-Ass-Data-Broker-Opt-Out-Li...
From a purely B2B perspective, the most egregious offender IMHO is Zoominfo largely because of the wide adoption in sales orgs. You can opt-out here: https://www.zoominfo.com/privacy-center/update/remove
This is one of the better ones I've seen: https://github.com/yaelwrites/Big-Ass-Data-Broker-Opt-Out-Li...
From a purely B2B perspective, the most egregious offender IMHO is Zoominfo largely because of the wide adoption in sales orgs. You can opt-out here: https://www.zoominfo.com/privacy-center/update/remove
Change your number and only leak to IRL contacts.
Optionally have your work provide you a work phone and only give that out for work activities.
Optionally have your work provide you a work phone and only give that out for work activities.
Also the same Datadog that in order to give us a price break due to a misconfiguration, strong-armed us into signing up for additional monitoring.
What kind of monitoring did they strong-arm you into? How did they strong arm? Genuine questions!
It's been long enough that I'd have to dig up the emails. May have been RUM, uncertain. Bottom line is we have never been treated that way by any other company. No matter how many podcasts they put ads on or events they sponsor at conferences, they are the tech equivalent of used cars salesmen; they are not friends of developers.
ASS - acronym’s seriously suck
Wish I could use this
Datadog is the Monster cable company of data analytics
cool brand name.
DRUIDS is the design system for Datadog. It stands for “Datadog Reusable User Interface Design System.”
Datadog is an application monitoring suite (distributed trace, log aggregation, infrastructure instrumentation, dashboarding) that includes a web-based front-end which makes use of these components.
Your explanation assumes people know what Datadog is, or that what that is could somehow be easily inferred from the name ...
It's one of the most prominent vendors in the tech space. It's pretty safe to assume people know who Datadog is on a software forum. Not always well loved, but definitely well known.
Absolutely not. In 20 years I've heard the name twice: Once six weeks ago when they emailed me to try to get my team to use their tools (which they failed to describe to me, so I declined), and a second time just now on HN.
I don't know who they are or what they do. I asked my coworkers - they didn't know either.
"Everybody knows about x" where x is any proper noun in the software space is frequently a bad bet. The software world is exceedingly large, and people are familiar with chunks of it.
I don't know who they are or what they do. I asked my coworkers - they didn't know either.
"Everybody knows about x" where x is any proper noun in the software space is frequently a bad bet. The software world is exceedingly large, and people are familiar with chunks of it.
...I'm really struggling to understand why you can't/won't figure out what DataDog does the same way any of us figures out what anything does; by Googling it.
You asked your coworkers! That demonstrates an interest, why wouldn't you ask Google?
You asked your coworkers! That demonstrates an interest, why wouldn't you ask Google?
There are always people who are out of touch with the current market (and I don't mean that condescendingly. There's not much value in knowing these things for many people in the space). That doesn't mean things they aren't aware of aren't well known. Of course not everyone knows, but a critical mass certainly does such that it's not really necessary to introduce the company every time it's discussed.
> Absolutely not. In 20 years I've heard the name twice: Once six weeks ago when they emailed me to try to get my team to use their tools (which they failed to describe to me, so I declined), and a second time just now on HN.
I want to second this - today is the second time I heard of them. The first was a job offer on LinkedIn that also failed to describe what Datadog did. It did talk a lot about how it was enterprise scale and important, though.
I declined to look further into it.
I want to second this - today is the second time I heard of them. The first was a job offer on LinkedIn that also failed to describe what Datadog did. It did talk a lot about how it was enterprise scale and important, though.
I declined to look further into it.
Well, it was only founded in 2010. So it's been less than 12 years in existence.
And one of the better known vendors in monitoring services.
And one of the better known vendors in monitoring services.
I know who they are only because I googled them while reading this thread. There are a lot of software developers who don't deploy software via the cloud, so would have zero use for "Cloud Monitoring as a service"
Never heard of it, if you're not a webshitter there's about a 0% chance you'd ever know who or what it is.
Honestly the first time I've ever heard of them.
even npm package[2] asks for login
[1] https://druids.datadoghq.com/foundations/contribute
[2] https://www.npmjs.com/package/@druids/ui