December 17, 2024 • 35 mins
What if you could scale your SaaS platforms effortlessly across diverse hosting services? Join us as we welcome Adam McCrea, the brilliant mind behind JudoScale, who takes us through his fascinating evolution from being a Rails developer to creating a cutting-edge autoscaling solution. Adam opens up about the technical challenges he faced while adapting JudoScale for platforms like Render, Fly, and Railway, and how Heroku's unique architecture initially shaped his approach. His journey is one of innovation driven by necessity, as JudoScale originated from a need to optimize costs more efficiently than existing solutions.
Our conversation doesn't shy away from complexity; in fact, it embraces it. Adam shares his experiences of grappling with AWS integration, navigating the intricate maze of ECS, EC2, Fargate, and IAM, all driven by customer demand. We explore the strategic shift from metered billing to flat-tiered pricing and the hurdles faced while setting up a staging environment on Render, ultimately reaffirming Heroku's smoother experience. This episode promises valuable insights into the strategic decisions and architectural reimaginations that keep JudoScale ahead of the game.
Adding a creative flair, we delve into the entertaining world of infomercial production, as Adam recounts his experience crafting a humorous Billy Mays-inspired ad for JudoScale. With the aid of AI tools like ChatGPT and Descript, Adam turned a fun concept into an engaging reality. As we wrap up, Adam shares his excitement for RailsConf in Philadelphia and the significance of fostering connections through digital networking. Whether you're a tech enthusiast or a developer seeking innovative scaling solutions, this episode is brimming with insightful takeaways and creative inspiration.
Honeybadger is an application health monitoring tool built by developers for developers.
Disclaimer: This post contains affiliate links. If you make a purchase, I may receive a commission at no extra cost to you.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Hello everyone, welcome to another episode of
Code and the Coding Coders whoCode it.
I'm your host, drew Bragg, andI'm joined today by Billy Mays
or Adam McRae I can't rememberwhich Adam.
For anyone who's not familiarwith you, would you please
introduce yourself?
Speaker 2 (00:14):
Yeah, so I'm Adam.
I created a SaaS app calledJudoScale in 2016.
It was originally RailsAutoscale.
We renamed it in 2021.
It's an autoscaling add-on,originally an autoscaling add-on
for Heroku, and now we'reexpanding to other platforms
like Render and Fly and Railway.
Yeah, I've been a Railsdeveloper since 2007.
(00:35):
So, yeah, I've kind of beenaround, done a lot of things and
landed on JudoScale, which I'vebeen doing full-time since 2021
.
Speaker 1 (00:43):
2007,.
That's pretty OG.
So the way this is going towork is I'm going to ask Adam
three questions.
I'm going to ask him what he'sworking on, what kind of
blockers he has.
If he doesn't have a currentblocker, he can talk about a
recent blocker he had and how hewent about solving it.
And then the last question, myfavorite, is what is something
cool, new or interesting thatyou've recently learned,
discovered?
Interesting that you'verecently learned, discovered?
It doesn't have to be codingrelated, but it can be.
(01:05):
So let's dive in, Adam, whatare you working on?
Speaker 2 (01:14):
Last year we did our first platform expansion outside
of Heroku.
We integrated with Render andwe integrated with Amazon ECS
and it works fine.
But we made some decisions thatI'm not entirely happy with.
So right now I'm peeling backsome technical debt.
Basically, when we originallybuilt Rails Autoscale, it was
all Heroku.
Everything in our app wasmodeled after Heroku stuff.
We got dynos, we got apps, wegot processes All the Heroku
terms, the Heroku concepts werelike baked in.
(01:37):
And then when we did render ourfirst expansion beyond Heroku,
that required a lot ofrefactoring, peeling back layers
of the onion and just findingall these places where it's
assumptions of heroku built in.
And one that made thatparticularly challenging with
render at the time is thatrender just has a flat list of
services, whereas heroku hasapps with processes in it.
(01:57):
You got your web and yourworker processes within an app.
Render, at the time at least,was just a flat list of services
.
There's no notion of an app.
So you got your web and workerservices all mixed in with all
your different other web andworker services, and the way we
ended up modeling that in Judoscale is kind of making it
difficult to now expand to otherplatforms.
(02:17):
So basically, I'm totallyre-architecting how we model our
render data within Judo scaleto make it easier to expand to
other platforms, because we'recurrently building integrations
with Railway and Fly.
So I'm working on a lot of kindof tech debt stuff right now.
Speaker 1 (02:33):
That's interesting.
The not just oh Heroku calls ita dyno and another company
might call it something else.
We need a universal term, butstraight architecture decisions
being influenced by what you'retrying to integrate with.
You can talk about how you'redoing it now, or you can talk
about how you did it with render, but are you almost treating
them like separate apps or justan entry point with a facade, so
(02:58):
that everything kind of worksthe same way, but under the hood
it's drastically different?
What was the approach thereWithin Judo?
Speaker 2 (03:04):
scale, there's always going to be a grouping of
processes and on Heroku thatgrouping is an app and the
processes are processes.
On something like Railway, theyhave notions of projects and
environments and within thoseenvironments you have services.
So in that case, your projectenvironments, your judo scale
(03:24):
production, that's yourenvironment and within there
you've got services.
It seems to work well that eachplatform we integrate with has
some kind of grouping for theirservices.
Like I said, render at firstdidn't, but now Render has very
similar to Railway.
They have projects andenvironments that you can group
your services into.
So them adding that I thinkthey added it last year, maybe
(03:47):
earlier this year is kind ofletting us be more consistent
with how we model this stuff inJudo scale.
Behind the scenes in Judo scalewe're still kind of using some
Heroku terms.
We've got apps and processesDepending on which platform
we've integrated with.
We'll just call it differentthings on the front end, because
each platform calls itdifferent things.
Speaker 1 (04:08):
Okay, I want to go back, probably a little further
than that.
How did the idea to build anautoscaler first come to you?
Was it just you needed it andyou built it, and then other
people were like that's cool, Iwant it, or was it?
No, I think this is definitelya business I could run.
Let me build this thing out Alittle of both.
Speaker 2 (04:27):
I was working at a small company at the time.
We basically generated leadsfor real estate agents and when
I joined they were running onEngine Yard.
I kind of spearheaded a movefrom Engine Yard to Heroku.
Because we were just paying aninsane amount for Engine Yard
for not a whole lot, that wewere getting Moved to Heroku,
because we were just paying aninsane amount for Engine Yard
for not a whole lot, that wewere getting Moved to Heroku and
(04:47):
it was great for us.
Everything worked really well.
We saved a lot of money in thatmove.
At the time we were runningeight dynos 24-7 for an app
that's primarily just usedduring the day.
It's sort of a back officesystem kind of thing.
I knew we wanted to autoscaleit to be able to avoid wasteful
spending on our hosting and wetried a couple of the auto
scalers that were available atthe time.
(05:09):
This is 2016.
And there were only a coupleand they just didn't work very
well.
They were painful to use, theywere confusing, they weren't
very reliable and I had someideas on how to build a better
auto scaler.
So I was like, what if I justdid this as a side project?
It sounds fun, we could use itat the company I was at.
If it works out, then maybe Ithrow it into the Heroku
(05:32):
marketplace as an add-on and seewhat happens.
So I asked my boss do you careif I do this as a side project
and I own the IP, but we use itif it works?
He doesn't care.
It was a super small company,very casual about all that stuff
, so I built it, we used itinternally and it did work great
.
So then I started integratingit as a Heroku add-on and that
launched in 2017.
(05:52):
And at that point I definitelyhad the dream of going full-time
on it.
But I wasn't the type of personthat, like I'm not going to
burn all my nights and weekendson this thing to make it happen
as soon as possible.
I was kind of content to justlet it grow slowly.
So I kept working on it, butnot at an exhausting pace or
anything like that, just kept itgoing as a side project where
(06:13):
it just kind of slowly grew inthe Heroku marketplace.
Speaker 1 (06:17):
When did it become kind of evident that, okay, this
is my full-time job, now I needto spend full-time on it to
keep it up, or was it more of a?
I can now spend full-time on it.
Speaker 2 (06:30):
Yeah, it's never been the kind of thing where it's
needed full-time attention.
Fortunately, it runs reallyreliably.
It does not get a lot ofsupport requests.
At this point, we've gotupwards of 800 customers and we
get maybe five support tickets aweek, something like that.
It's a really low maintenance,low support app, which has been
(06:51):
wonderful.
So, yeah, 2021, that was theyear that I went full time on it
.
Building up to that, 2020 was agood year.
It grew pretty well.
I had a number in my mind whereI wanted to hit where I'd be
comfortable going full-time onit.
At the time, my wife wasn'tworking, she was back in school,
so I was the only income.
And the problem is lifestylecreep is a real thing and I had
(07:16):
been a software developer for along time, so I was making
pretty good money as a softwaredeveloper.
I can't make the leap tofull-time until this product is
making pretty good money, if Idon't want to tell the family
hey, look, all these thingswe're used to, we're going to
cut back and I didn't want to dothat.
So, yeah, 2021, two thingshappened at the same time.
Revenue hit a point where Iwould have been comfortable
(07:37):
going full-time on it, but itwould have felt like a little
bit of a risk.
And at that time I actuallyapplied to TinySeed, which is,
imagine, a Y Combinatoraccelerator, but on a much, much
smaller scale, for bootstrapcompanies, so companies with one
or two co-founders that want tobuild a company that is
profitable and want to grow itbut not grow at all costs like a
(08:02):
VC funded thing.
So, tinyseed the way it worksis they give you a little bit of
funding I think they're publicabout it it's 120K for solo
founders, 180k for multiplefounders and they take a little
bit of equity for that.
And then you're in a programwhere you have mentorship and
you have community and they helpyou out.
And so I got accepted intoTinySeed, which gave a little
(08:25):
cash infusion, gave completeconfidence that I could go
full-time on it with almost norisk.
And so, yeah, that all happenedin 2021, made the leap to
full-time on it and, yeah, didthe company need my full-time
attention?
It didn't.
It was kind of my dream that Iwanted to be doing my thing
full-time, so I was eager to dothat.
Speaker 1 (08:43):
That's awesome, and is that kind of being able to
work on it full-time because ofTinySeed is that what drove you
to be like, okay, where else canwe stick this?
Or was that a?
There's a lot of people whowere like I don't want to be on
Heroku anymore, I want to beover here.
Can I please use your service?
Speaker 2 (09:00):
The customers who were leaving judo scale.
When we reached out to them andthey replied, the usual
response was oh, we love judoscale, we're leaving heroku and
we're going to.
At the time it was aws or wasrender all right.
Well, these people want to keepusing judo scale, so let's
build something so that they can.
So that was what motivatedthose.
Since then, we've had more andmore customers asking about
(09:22):
railway and fly, so that'sbasically what drives these
decisions are what our customersare asking for, and usually
they're not asking for newfeatures in the app.
Usually it's a handful who areleaving Heroku, want to go to a
different cloud provider andwant to be able to keep using
Judo Scale.
Speaker 1 (09:38):
The biggest challenge with the move is how everybody
names and architects theirparticular services.
We've already discussed howmuch learning did you have to do
going to render.
Is it once you figure out theirnaming scheme and setup,
scaling up and down andinfrastructure is basically the
same across all, or is thereeven more layers of complexity
(10:01):
once you get past naming andorganization?
Speaker 2 (10:04):
It's pretty consistent.
They all work a little bitdifferently in a few different
ways, but at the end of the day,you're deploying containers and
scaling them up and down andit's not that difficult to
migrate between them and build amental model between them.
With a big asterisk, the onethat was a pain in the ass,
unsurprisingly, was AWS.
Speaker 1 (10:23):
So, shocking.
Speaker 2 (10:25):
I avoided that for a long time and eventually just
had enough customers who hadleft Heroku to AWS and wanted to
use JudoSkill.
I'm like this is stupid, thatwe're not taking the time to
investigate this and buildsomething.
So, yeah, that was morechallenging because I didn't
really have any experience withAWS beyond, like creating an S3
bucket.
That's the extent of what I haddone in AWS at that point.
(10:48):
So, yeah, kind of had to learnhow ECS worked and the
difference between using it withEC2 and Fargate, all the
acronym soup that's in AWS.
Oh man, the permissions withIAM, all that stuff.
That was a painful integrationfor sure, and I still dread
anytime I have to get into AWSfor anything.
Speaker 1 (11:10):
Is it to the point where you're like, okay, that's
done, I can mostly just let itsit there.
Or is it?
No, there's still a bunch ofstuff to do, but it's good
enough for now.
I can just touch it when I'mfeeling.
Pain is going to be fun today,I don't know.
Speaker 2 (11:22):
We've got a list of quality of life improvements we
want to make on that platformand, more generically, on sort
of all the non-Horoku platformsthat we've integrated with.
It just works and we don'treally have to dig into it for
the most part Just a lot oflittle improvements that we want
to make.
And then the biggest changethat is on our radar that I
either want to make this monthor next month is redoing our
(11:46):
pricing for the non-Horokuplatforms On Heroku.
Because we're in themarketplace, we're pretty much
tied down.
We have to do monthly tieredfixed price billing and when we
did our rendering, ECSintegration made the decision to
go with metered billing basedon how much compute usage only
because I couldn't think of agood way to do the flat tiered
(12:08):
billing because of the way ECSand render work.
But it's been a pain.
Meter billing is such a pain Iwant to throw it out and we're
going to throw it out.
We're going to migrate to justflat tiered pricing.
It'll make our life easier.
It'll make our customers' liveseasier.
It'll make it easier tounderstand what your price is
going to be before you sign upfor it Another one of those
(12:29):
decisions that I'm having toundo from the initial render and
ECS integration I feel yourpain.
Speaker 1 (12:36):
Billing anytime is scary and difficult Figuring out
how to do billing and then,even whenever you implement
anything, any change feelsdangerous because you're dealing
with people's money.
It's terrifying.
Speaker 2 (12:50):
Yeah, that was one of the best things about building
on Heroku is that we didn't haveto think about billing.
Heroku handles all the billing.
They charge a premium for it.
It's an app store like Applewhere they take a 30% cut.
But the reality is I wouldn'tbe doing this full time if it
weren't for Heroku's add-onmarketplace, because I was able
to just throw it in there and,with almost no marketing, people
(13:12):
would just find it and sign upfor it and start using it.
So it just kind of grew slowlyon its own.
So I've never really had aproblem paying that tax to be in
the Heroku marketplace.
But yeah, when we expandedbeyond that, yeah, we had to
handle the billing and so thatwas a whole new world and it is
super stressful.
Speaker 1 (13:31):
So, now that you have to have billing outside of the
heroku marketplace, is the ideato?
Maybe it'll still be in theheroku marketplace?
Because why not?
But also we would prefer if youdidn't for heroku customers.
Speaker 2 (13:44):
We're definitely still going to be in the
marketplace and we'll stillencourage that they go through
the marketplace.
But, yeah, for anybody who'snot on heroku, obviously, when
it comes to which hostingplatform to use, I'm still a fan
of Heroku.
I still think Heroku is a greatchoice and we're still hosted
on Heroku mostly.
Yeah, we actually just movedour staging environment to
(14:06):
render a month ago, but we'reabout to move it back.
That's a whole nother thing.
But, yeah, definitely still afan of Heroku.
Speaker 1 (14:12):
If anybody's using Heroku, we're going to encourage
them to go through themarketplace.
We're not going to not diveinto staging on render, but not
for much longer.
Was that an experiment?
Was that just hey, let's seehow render works.
What was that move for and whyare you moving back?
Speaker 2 (14:28):
It was sort of all of the above.
It was my own curiosity.
I wanted to see what's it likerunning an actual product on
render.
I wanted to see, cost-wise, howis that going to compare?
Because our staging environmentwe call it a staging
environment, but it's a veryproduction-like environment Our
staging app actually auto-scalesour production app and our
(14:48):
production app auto-scales ourstaging app, so our staging app
still gets pretty decent traffic.
So I wanted to see how the costwould compare.
I wanted to see how theexperience would compare and I'm
trying to do more video contentand that kind of thing and I
thought this would be a goodopportunity for some video
content to show what's it liketo migrate an app from Heroku to
Render and then do somecomparison there.
(15:09):
But yeah, what's beeninteresting?
After leaving it there for amonth, I actually just got our
first full month bill for theRender app and it's crazy,
looking at the compute a lotcheaper than Heroku.
But what was surprising isRender charges you for egress,
(15:30):
so the outgoing bandwidth thatHeroku doesn't charge you for,
and that was 80% of our bill onrender and that's something we
didn't really account for whenwe experimented with the move
and I'm like, yeah, if we moveour product to render, it's
going to be more expensive thanHeroku.
So that, amongst a few other,just little bumps along the road
(15:50):
that just make Heroku asmoother experience.
Speaker 1 (15:55):
And I mean, if you're more comfortable and familiar
with it, why not right?
If it's not saving money orsaving headache or anything, why
bother?
Speaker 2 (16:03):
Yeah, but I definitely want to be able to
serve our customers better.
I want to be even more familiarand comfortable with all the
platforms that we support, soI'm going to experiment with
moving to Railway and fly, justto see what it's like.
Speaker 1 (16:16):
Yeah, it makes sense.
Hey, these are the products weoffer and here's if you're
trying to move between them oranything, having that video
content is always a nice thing.
Speaker 2 (16:25):
I appreciate video content like that we got a lot
of people asking us for adviceon how to choose between them,
and it's hard because at the endof the day they all do the job
well enough and it's kind oflike try them out and see which
one feels the best.
It's a lot of work to try themout all individually, so maybe
(16:47):
if I can try them all out andproduce enough content to help
out with that, yeah, even whatyou learned with render is oh
yeah, the compute's about thesame, so that's about the same.
Speaker 1 (16:56):
But if you have a high Ingress app, that's where
Render gets expensive and maybeother people wouldn't know that.
And that's a what is your app,what is your app doing, kind of
follow-up question where westart getting into.
Maybe there is one that isbetter than the other for
certain types of apps.
Speaker 2 (17:14):
Yeah, one of the things that I want to build in
the next month or two is apricing calculator that covers
all of the popular pass servicesHeroku, render, fly, railway
because they all charge sodifferently.
They all have their own pricingcalculators that work
differently.
Some of them charge for egress,some of them don't.
Some of them charge for teammembers, some of them don't.
(17:35):
Some of them charge for teammembers, some of them don't.
Especially as somebody who isstill a fan of Heroku and sees a
lot of the hate that Herokugets on social media and whatnot
, it's not as clear cut as youthink.
Yeah, heroku does have somedynos per fem that are
ridiculously expensive and youshould basically never use, but
(17:55):
it doesn't mean that it's alwaysgoing to be astronomically more
expensive than the alternatives.
Speaker 1 (18:00):
Hidden costs and fees , and Ticketmaster springs to
mind.
Speaker 2 (18:05):
Yeah, I'm not going to label any of them with
Ticketmaster.
That's brutal.
Speaker 1 (18:09):
So what kind of blockers do you currently have?
Or, if there isn't a blockerthat you're currently dealing
with, what's a recent one thatyou had and what did you do to
get yourself around it or over?
Speaker 2 (18:19):
it need to put some time and energy into marketing.
So just making that time formarketing and forcing myself to
do the hard work of marketing,that's just been a continual
(18:41):
challenge.
This year I hired my firstfull-time developer, carlos
Antonio, who's a Rails corecontributor, and I worked with
him at a previous job.
He's working with me full-timenow.
So that has helped me put somemore time into marketing because
I know he is going to keepmoving the product forward even
when I'm not Even with that.
(19:04):
It's hard, because I reallyreally like working on the
product.
There's a lot of things that Iwant to get done.
One thing that I'm startingjust this week actually, is a
couple of friends of mine,colleen Schnettler and Leanna
Patch.
They are starting a new thingthey're calling SaaS Marketing
Gym, where I think it's mostlykind of solo founders like me or
(19:26):
similar get together for anhour a day.
It's like going to the gym.
It's like an hour set asidewhere we're coming together, we
have our goals in marketing andwe're just going to focus on our
marketing tasks and Leanna andColleen are going to be there to
help us get on, stock andbounce ideas off of and stuff
like that.
So actually doing my first gymsession right after we finished
(19:47):
recording this podcast.
Yeah, I'm super excited aboutthat.
I think that's going to help mekeep moving the ball forward on
the marketing side of things.
Speaker 1 (19:56):
Yeah, that's interesting.
I've talked to a few people nowrunning their own SaaS or
either as a solo or co-foundersetup, but that seems to be the
blocker.
I don't know how to market.
I don't know how I'm supposedto drive up people knowing about
my product and without justtalking about it 24-7, which
(20:19):
feels weird.
If I ever came up with a goodenough idea to run with, that
would be my big blocker too.
I don't know how to tell peopleabout my thing that I'm doing,
and I wouldn't even know whereto go to learn.
You want to go learn how tocode?
There's a number of differentplaces that will teach you how
to code.
Do those places exist formarketing, like a marketing
(20:39):
bootcamp?
Here's how to talk to people.
Here's why this phrasing worksbetter than this phrasing.
I run into it with designers.
Bump up that line by one pixeland use this font instead.
The entire page changes Likehow the heck did you see that?
Speaker 2 (20:54):
I can't see that until it's done, that is a big
reason that I joined TinySeed isto get community and mentorship
around marketing, because I didfeel that was my huge gap, and
part of it is knowing what to do, and the other part is just
making yourself do it.
Because, as a developer and Iknow a lot of developers can
sympathize with us I don't likebeing marketed to, so I don't
(21:17):
want to be the person who'salways talking about my product,
who's always pushing it onpeople.
Obviously, there are tastefulways to do it, there are
authentic ways to do it, and sothat's the challenge is how to
promote your product, buildawareness around your product,
but in an authentic way thatyou're genuinely trying to help
people and contribute to theoverall community, rather than
(21:41):
just tell people to use yourthing.
Speaker 1 (21:43):
Well, I think your first judo scale commercial was
a huge hit.
I loved it at the very least.
I got to ask how did you comeup with that?
Speaker 2 (21:53):
You tweeted that you were looking for companies to
sponsor prizes for the game showat RubyConf and I reached out
to you.
I was like, yeah, we'd love tosponsor some prizes.
And that's when you threw outthat, yeah, I've got extra time
this year, so if you want torecord a 30 second commercial,
I'll show it on stage.
The only requirement is that ithas to be funny.
And I was just oh well, I'm notgoing to do that.
(22:15):
That's way too much pressure.
I don't consider myself a funnyperson Like.
That's not me.
I was talking to a buddy of minewho actually helps me out on
Judo scale.
So, in addition to Carlos,who's full-time on Judo scale a
friend of mine, local to centralOhio where I live, he spends a
few hours a week with me on judoschool.
(22:35):
We get together at a coworkingspace and just kind of pair on
things and bounce ideas off ofeach other.
I was talking to him about it.
One of the reasons I loveworking with him is he always
pushes me out of my comfort zone.
He always challenges me, and soI told him about this.
This would be a hugedistraction.
I don't even know where we'dbegin.
He's like oh, we're definitelydoing it.
We're definitely doing it, yes.
(22:55):
So yeah, he and I just went outto lunch and started throwing
some ideas around.
Yeah, he was the one who came upwith the Billy Mays idea and
then we actually used ChatGPT towrite the initial script for it
.
We had the idea of let's havean actual scale.
The actual scale is the judoscale and that's what we're
(23:19):
selling in this infomercial.
So we kind of described that toChatGPT and it wrote the
initial script, nailed theinfomercial Billy Mays tone and
everything.
We tweaked it a ton but itreally helped get us past the
initial blank page.
Where do we start?
It gave us a starting point.
We iterated on that.
Once we felt good with thescript, timed it and got it to
(23:41):
that 30 second mark, we gottogether in his kitchen and
recorded it in a couple hoursand I spent maybe a day editing
it.
Probably didn't even need tospend that much.
I was just having way too muchfun adding sound effects and
making the cuts and everything.
It was so much fun.
Speaker 1 (23:58):
It turned out so great.
I had that idea when I was atRailsworld I was talking to Josh
Wood about a new Honeybadgerwas going to be sponsoring,
because they sponsor the podcastnow and I just, man, when I
first first did this whole gameshow thing, I thought about
having a commercial.
When I first first did thiswhole game show thing, I thought
about having like commercial.
I don't know if you've everseen any of wise old conference
(24:22):
talks, but he was just super outthere and he would throw random
commercials in his talks.
It wasn't a commercial, but itwas like a video that came
across like a commercial, but itwould talk about Ruby stuff.
Man, it'd be so funny to docommercials or something, and
he's.
I think we could do commercials.
Why not?
If you've got the hour now fora keynote, you could if you
wanted to, and I was all right.
Well, if you give me a 20 to 30second commercial that's funny
Then I'll roll with it and I'lltalk to anyone else, offer it up
(24:45):
, and I sent it to you too, andthis is great way of being like
thank you for sponsoring theprizes and also adding to the
game show atmosphere.
And, man, when I saw yours, Idon't know if I could even play
this on stage, because I can'tstop laughing.
It was so good, it's so good.
I'm going to link it when Ipost about this.
(25:07):
I think it's on your Blue Skyaccount too.
You posted the full video, butI'll make sure it's in my post
about this episode, because itis just so, so good.
Speaker 2 (25:18):
Thank you, I appreciate it.
We got a lot of good feedbackat the conference.
The commercial didn't evenreally explain what Judo Scale
does.
Speaker 1 (25:25):
Whatever, it's fun.
It got people to Google JudoScale, I guess, so okay.
So the last question other thanyour new career in infomercial
production, what is somethingcool, new or interesting that
you've recently learned,discovered?
Got into hobby-wise, code-wise,doesn't matter.
What is something you want toshare?
Speaker 2 (25:47):
It actually is related to the commercial,
because I have been loving usingDescript for making video
content.
I've been trying to make somemore video content, whether it's
for Blue Sky or LinkedIn orYouTube or a RubyConf commercial
, and if you want to get intovideo content at all, I can't
think of any reason not to useDescript.
(26:09):
I'm not getting paid by them,it's just phenomenal.
So Descript is a video editingapp.
You record in it, ittranscribes your video and then
you edit the script of the videoand that edits the video for
you.
You cut out words from thescript.
It cuts out that part of thevideo.
It actually does a really goodjob with the edit points too.
Sometimes I have to tweak them,but most of the time I don't.
(26:31):
Really good job with the editpoints too.
Sometimes I have to tweak them,but most of the time I don't.
It can do the captions, it cando highlight reels.
It can do so much.
So yeah, if you want to getinto video content, you've got
to give Descript a try.
Speaker 1 (26:41):
It makes it so easy?
Yeah, and I'll expand on that.
If you are a new podcaster andyou are not ready to have your
own editor, but you want toclean up the audio feed, I used
the script in the beginning too.
It's really easy to reducepauses.
It'll automatically take outuhs and ums and all those filler
words.
Yeah, it also could edit thescript right there.
So you're like oh, I have toremember to take out that
(27:03):
bumbling run-on sentence that Ihad there.
You could just delete it and itwill delete it for you.
Speaker 2 (27:08):
You might tweak the audio a little bit, but yeah
it's a very cool application andit even has a library of sound
effects and things like thesound effects that I put in the
commercial.
Those were in Descript's soundeffects library that I was just
able to throw in there.
It was awesome.
Speaker 1 (27:23):
That's a good one.
Yeah, I think that Descript isa wonderful product and has a
lot of uses Video.
I've used it for internalvideos when I was at within
three and, like we would gothrough refactoring, I would
just record it.
It was so cool because I don'tknow if I can do a video this is
going to be crazy.
But I was using Descript to dopodcasting already and I was all
right well, try their videofeature and it made it so dummy
(27:45):
easy to do what I was trying todo to do.
I was very impressed by thefinal product to the point where
I was wow, maybe this is athing, and then it's not for me,
but it's way too much work andI'm entirely too much of a
perfectionist to do that.
But yeah, it's such a cool tool.
I believe my podcast editor,paul, who does a bunch of the
other podcasts, edits inDescript also, so it is
(28:09):
definitely a tool for theprofessionals too.
Speaker 2 (28:11):
Yeah, and the other thing I would add is, if anybody
is kind of curious about doingvideo content but is intimidated
by it at all, especially ifyou're thinking about needing to
get specialized equipment orwhatever.
The other thing that I wouldrecommend, if you have a modern
iPhone and a modern Mac, youriPhone is an excellent webcam.
I don't know if you publishedthe video for these things.
(28:31):
The camera that I'm using rightnow is my iPhone, just on a
magnet clip at the top of mymonitor, and, yeah, continuity
mode is built into the newerMacs and I think it looks great.
Obviously, you can get DSLR orwhatever and get marginally
better video, but I think thestep up from a traditional
webcam to using your iPhone as awebcam is huge and if you
(28:54):
already have those things, youcan make pretty good videos with
minimal equipment.
You got to get a good mic and amic arm.
That's the biggest thing.
And, yeah, then use your iPhone, use Descript, you can make
good stuff.
Speaker 1 (29:06):
I don't publish the videos.
I've thought about it whenYouTube started hosting podcasts
, but no one wants to see thisface.
I don't want to see this facethat much.
So I try not to look in themirror too often, even, but to
piggyback on the.
You can make minor improvements.
Even lighting makes a hugedifference.
I have some cool lights.
I can turn them off.
Video is still fine, but to meit gets much better once you
(29:31):
introduce some really goodlighting too.
So for sure, you know, itdoesn't have to be the most
spectacular camera I definitelydon't have a very fancy camera
but the lighting makes a hugedifference.
I even run the lighting justwhen I'm on a zoom call, just
because it looks marginallybetter.
So why not?
So yeah, you can use youriphone, you can use some extra
(29:52):
lighting.
Doesn't have to be anythingfancy, just a lamp with some
more light will do.
But yeah, I think themicrophone, the audio, is in
many ways the more important bitof even a video, because you
got to hear what's going on.
So is there anything else thatyou wanted to talk about?
Touch on, discuss before wewrap.
Speaker 2 (30:13):
I'm talking about RailsConf.
How are you feeling aboutRailsConf in Philadelphia?
Does that excite you at all?
I'm so excited.
Speaker 1 (30:19):
Could you tell at RubyConf me bouncing around like
an over-caffeinated squirrel.
I am pumped.
We've never had a Ruby orRailsConf in Philadelphia, which
is horrible in my opinion.
Railsconf in Philadelphia,which is horrible in my opinion.
But that will all change inJuly when RailsConf comes to
Philly.
The very last RailsConf will bein Philly.
(30:39):
I'm super pumped.
I'm actually going down to thecity this weekend for a
different event, but it'll be inthe same area of operation as
where the conference will be.
So I'm going to do a little bitmore than just my mindless
walking around and just, oh,that smells good.
I'll go in this restaurant.
I'm going to kind of hit someplaces that I know I want to get
good information on and sendthat back to Ruby Central to try
(31:01):
and help make a visitor's guideand the big things like
Independence Hall and theLiberty Bell People know about
that.
But what else is there in Philly?
What's surrounding Philly?
In case you're coming out for alonger period of time and
you're bringing the familybecause it's in July instead of
a weird in school month thistime there's a lot of cool stuff
around Philly, in Philly.
So I'm very excited to do thatkind of stuff, see how many
(31:23):
people come out and go.
Yeah, you're right, philly isreally cool.
I'm glad we had this experiencehere.
But, yeah, great food.
The area that the conferencewill be in is a lot of fun, so
very, very excited.
Speaker 2 (31:35):
I'm excited too.
It's going to be the firstconference that we'll be an
official sponsor of.
We're going to have a booth andhoping to get the whole
JudoSkill team out there allthree of us.
We're talking about beinguncomfortable doing the
marketing stuff.
The idea of standing at a boothand having to talk to people
about JudoSkill as they walk upto it is absolutely terrifying
to me.
It's time.
It's time to force myself outof my comfort zone and do that.
Speaker 1 (31:59):
You can also take a page out of Honey Badger's book.
I think Ben and Josh do a greatjob.
They just have this booth withcouches and they're there and
they're chilling and maybethey're hacking on something
with someone and they've got abanner and if someone comes up
and they go, hey, what's HoneyBadger?
They'll get up and they'll talkto them, but otherwise it's
just a cool space to hang outand chill and I feel like
everybody knows what HoneyBadger is.
Oh, definitely that kind ofpassive marketing in a way at a
(32:25):
place like RailsConf can beawesome and you'll have a TV
playing your Billy Mayscommercial.
Speaker 2 (32:31):
There's only so many honey badger ideas that I can
steal before it just becomesdisgusting.
So I love their lounge wherecome, hang out and chill.
We'll have to think ofsomething unique to do with our
booth.
Speaker 1 (32:38):
What's the swag situation going to be like?
What are you doing?
Don't know.
Yet I start brainstorming.
That's the key.
That's how you get them.
When you give them cool stuff,they'll remember you forever.
Speaker 2 (32:47):
My one rule is whatever swag we bring, it's
going to be a one size fits all.
We're not bringing differentsizes of t-shirts.
Figure, how many of each sizedo you need to bring?
Like no, it's just going to bewhatever, it is One size fits
all.
Speaker 1 (33:02):
Don't want an extra suitcase like Josh brings of
just a ton of t-shirts in agiant rolling case, because I
really bring a million of them,Although I will say you do have
very cool t-shirts.
I love the auto scale and chillt-shirt.
They're very soft andcomfortable and they come in a
variety of colors.
But the auto scale and chill isfunny to me.
It makes sense and it's a verygood t-shirt.
Speaker 2 (33:24):
I did an initial run of rails auto scale t-shirts
back in the day, had them madein a bunch of different sizes
and stuff like that, and then itended up just with a closet
full of t-shirts.
And then we rebranded and stillhave a closet full of t-shirts.
I decided I wasn't going tomake that choice again.
So, yeah, when we made the judoscale shirts, we're going to do
them print on demand as peoplewant one, which, yeah, if you're
(33:44):
listening to this and you wanta judo scale shirt, shoot me an
email, adam, at judoscalecom.
Speaker 1 (33:49):
I'll get you a t-shirt.
I recommend them.
They're comfy and hilarious.
So I was wearing mine atRubyConf and I will be wearing
mine at RailsConf.
Well, cool man, how can peoplefind you on the internet?
Yeah, I'm.
Speaker 2 (34:01):
AdamLogic on BlueSky.
Adamlogic on LinkedIn.
We are judoscalecom.
If you're interested inauto-scaling your web app, yeah,
I think that's it Wherever andauto-scaling your web app yeah,
I think that's it.
Speaker 1 (34:11):
Wherever it may be hosted.
Adam at JudoScalecom.
Speaker 2 (34:13):
I'm not covering that hosting platform.
I'm still old school and loveemail.
Shoot me an email,adamatjudoscalecom.
I'd love to hear from you.
Awesome, awesome.
Speaker 1 (34:20):
Billy.
Speaker 2 (34:21):
Oh yeah.
Speaker 1 (34:22):
So excited Yay to chat with us.
Thank you.
Speaker 2 (34:25):
Drew, this was a lot of.
Hello everyone, welcome to another episode of
Code and the Coding Coders whoCode it.
I'm your host, drew Bragg, andI'm joined today by Billy Mays
or Adam McRae I can't rememberwhich Adam.
For anyone who's not familiarwith you, would you please
introduce yourself?
Speaker 2 (00:14):
Yeah, so I'm Adam.
I created a SaaS app calledJudoScale in 2016.
It was originally RailsAutoscale.
We renamed it in 2021.
It's an autoscaling add-on,originally an autoscaling add-on
for Heroku, and now we'reexpanding to other platforms
like Render and Fly and Railway.
Yeah, I've been a Railsdeveloper since 2007.
(00:35):
So, yeah, I've kind of beenaround, done a lot of things and
landed on JudoScale, which I'vebeen doing full-time since 2021
.
Speaker 1 (00:43):
2007,.
That's pretty OG.
So the way this is going towork is I'm going to ask Adam
three questions.
I'm going to ask him what he'sworking on, what kind of
blockers he has.
If he doesn't have a currentblocker, he can talk about a
recent blocker he had and how hewent about solving it.
And then the last question, myfavorite, is what is something
cool, new or interesting thatyou've recently learned,
discovered?
Interesting that you'verecently learned, discovered?
It doesn't have to be codingrelated, but it can be.
(01:05):
So let's dive in, Adam, whatare you working on?
Speaker 2 (01:14):
Last year we did our first platform expansion outside
of Heroku.
We integrated with Render andwe integrated with Amazon ECS
and it works fine.
But we made some decisions thatI'm not entirely happy with.
So right now I'm peeling backsome technical debt.
Basically, when we originallybuilt Rails Autoscale, it was
all Heroku.
Everything in our app wasmodeled after Heroku stuff.
We got dynos, we got apps, wegot processes All the Heroku
terms, the Heroku concepts werelike baked in.
(01:37):
And then when we did render ourfirst expansion beyond Heroku,
that required a lot ofrefactoring, peeling back layers
of the onion and just findingall these places where it's
assumptions of heroku built in.
And one that made thatparticularly challenging with
render at the time is thatrender just has a flat list of
services, whereas heroku hasapps with processes in it.
(01:57):
You got your web and yourworker processes within an app.
Render, at the time at least,was just a flat list of services
.
There's no notion of an app.
So you got your web and workerservices all mixed in with all
your different other web andworker services, and the way we
ended up modeling that in Judoscale is kind of making it
difficult to now expand to otherplatforms.
(02:17):
So basically, I'm totallyre-architecting how we model our
render data within Judo scaleto make it easier to expand to
other platforms, because we'recurrently building integrations
with Railway and Fly.
So I'm working on a lot of kindof tech debt stuff right now.
Speaker 1 (02:33):
That's interesting.
The not just oh Heroku calls ita dyno and another company
might call it something else.
We need a universal term, butstraight architecture decisions
being influenced by what you'retrying to integrate with.
You can talk about how you'redoing it now, or you can talk
about how you did it with render, but are you almost treating
them like separate apps or justan entry point with a facade, so
(02:58):
that everything kind of worksthe same way, but under the hood
it's drastically different?
What was the approach thereWithin Judo?
Speaker 2 (03:04):
scale, there's always going to be a grouping of
processes and on Heroku thatgrouping is an app and the
processes are processes.
On something like Railway, theyhave notions of projects and
environments and within thoseenvironments you have services.
So in that case, your projectenvironments, your judo scale
(03:24):
production, that's yourenvironment and within there
you've got services.
It seems to work well that eachplatform we integrate with has
some kind of grouping for theirservices.
Like I said, render at firstdidn't, but now Render has very
similar to Railway.
They have projects andenvironments that you can group
your services into.
So them adding that I thinkthey added it last year, maybe
(03:47):
earlier this year is kind ofletting us be more consistent
with how we model this stuff inJudo scale.
Behind the scenes in Judo scalewe're still kind of using some
Heroku terms.
We've got apps and processesDepending on which platform
we've integrated with.
We'll just call it differentthings on the front end, because
each platform calls itdifferent things.
Speaker 1 (04:08):
Okay, I want to go back, probably a little further
than that.
How did the idea to build anautoscaler first come to you?
Was it just you needed it andyou built it, and then other
people were like that's cool, Iwant it, or was it?
No, I think this is definitelya business I could run.
Let me build this thing out Alittle of both.
Speaker 2 (04:27):
I was working at a small company at the time.
We basically generated leadsfor real estate agents and when
I joined they were running onEngine Yard.
I kind of spearheaded a movefrom Engine Yard to Heroku.
Because we were just paying aninsane amount for Engine Yard
for not a whole lot, that wewere getting Moved to Heroku,
because we were just paying aninsane amount for Engine Yard
for not a whole lot, that wewere getting Moved to Heroku and
(04:47):
it was great for us.
Everything worked really well.
We saved a lot of money in thatmove.
At the time we were runningeight dynos 24-7 for an app
that's primarily just usedduring the day.
It's sort of a back officesystem kind of thing.
I knew we wanted to autoscaleit to be able to avoid wasteful
spending on our hosting and wetried a couple of the auto
scalers that were available atthe time.
(05:09):
This is 2016.
And there were only a coupleand they just didn't work very
well.
They were painful to use, theywere confusing, they weren't
very reliable and I had someideas on how to build a better
auto scaler.
So I was like, what if I justdid this as a side project?
It sounds fun, we could use itat the company I was at.
If it works out, then maybe Ithrow it into the Heroku
(05:32):
marketplace as an add-on and seewhat happens.
So I asked my boss do you careif I do this as a side project
and I own the IP, but we use itif it works?
He doesn't care.
It was a super small company,very casual about all that stuff
, so I built it, we used itinternally and it did work great
.
So then I started integratingit as a Heroku add-on and that
launched in 2017.
(05:52):
And at that point I definitelyhad the dream of going full-time
on it.
But I wasn't the type of personthat, like I'm not going to
burn all my nights and weekendson this thing to make it happen
as soon as possible.
I was kind of content to justlet it grow slowly.
So I kept working on it, butnot at an exhausting pace or
anything like that, just kept itgoing as a side project where
(06:13):
it just kind of slowly grew inthe Heroku marketplace.
Speaker 1 (06:17):
When did it become kind of evident that, okay, this
is my full-time job, now I needto spend full-time on it to
keep it up, or was it more of a?
I can now spend full-time on it.
Speaker 2 (06:30):
Yeah, it's never been the kind of thing where it's
needed full-time attention.
Fortunately, it runs reallyreliably.
It does not get a lot ofsupport requests.
At this point, we've gotupwards of 800 customers and we
get maybe five support tickets aweek, something like that.
It's a really low maintenance,low support app, which has been
(06:51):
wonderful.
So, yeah, 2021, that was theyear that I went full time on it
.
Building up to that, 2020 was agood year.
It grew pretty well.
I had a number in my mind whereI wanted to hit where I'd be
comfortable going full-time onit.
At the time, my wife wasn'tworking, she was back in school,
so I was the only income.
And the problem is lifestylecreep is a real thing and I had
(07:16):
been a software developer for along time, so I was making
pretty good money as a softwaredeveloper.
I can't make the leap tofull-time until this product is
making pretty good money, if Idon't want to tell the family
hey, look, all these thingswe're used to, we're going to
cut back and I didn't want to dothat.
So, yeah, 2021, two thingshappened at the same time.
Revenue hit a point where Iwould have been comfortable
(07:37):
going full-time on it, but itwould have felt like a little
bit of a risk.
And at that time I actuallyapplied to TinySeed, which is,
imagine, a Y Combinatoraccelerator, but on a much, much
smaller scale, for bootstrapcompanies, so companies with one
or two co-founders that want tobuild a company that is
profitable and want to grow itbut not grow at all costs like a
(08:02):
VC funded thing.
So, tinyseed the way it worksis they give you a little bit of
funding I think they're publicabout it it's 120K for solo
founders, 180k for multiplefounders and they take a little
bit of equity for that.
And then you're in a programwhere you have mentorship and
you have community and they helpyou out.
And so I got accepted intoTinySeed, which gave a little
(08:25):
cash infusion, gave completeconfidence that I could go
full-time on it with almost norisk.
And so, yeah, that all happenedin 2021, made the leap to
full-time on it and, yeah, didthe company need my full-time
attention?
It didn't.
It was kind of my dream that Iwanted to be doing my thing
full-time, so I was eager to dothat.
Speaker 1 (08:43):
That's awesome, and is that kind of being able to
work on it full-time because ofTinySeed is that what drove you
to be like, okay, where else canwe stick this?
Or was that a?
There's a lot of people whowere like I don't want to be on
Heroku anymore, I want to beover here.
Can I please use your service?
Speaker 2 (09:00):
The customers who were leaving judo scale.
When we reached out to them andthey replied, the usual
response was oh, we love judoscale, we're leaving heroku and
we're going to.
At the time it was aws or wasrender all right.
Well, these people want to keepusing judo scale, so let's
build something so that they can.
So that was what motivatedthose.
Since then, we've had more andmore customers asking about
(09:22):
railway and fly, so that'sbasically what drives these
decisions are what our customersare asking for, and usually
they're not asking for newfeatures in the app.
Usually it's a handful who areleaving Heroku, want to go to a
different cloud provider andwant to be able to keep using
Judo Scale.
Speaker 1 (09:38):
The biggest challenge with the move is how everybody
names and architects theirparticular services.
We've already discussed howmuch learning did you have to do
going to render.
Is it once you figure out theirnaming scheme and setup,
scaling up and down andinfrastructure is basically the
same across all, or is thereeven more layers of complexity
(10:01):
once you get past naming andorganization?
Speaker 2 (10:04):
It's pretty consistent.
They all work a little bitdifferently in a few different
ways, but at the end of the day,you're deploying containers and
scaling them up and down andit's not that difficult to
migrate between them and build amental model between them.
With a big asterisk, the onethat was a pain in the ass,
unsurprisingly, was AWS.
Speaker 1 (10:23):
So, shocking.
Speaker 2 (10:25):
I avoided that for a long time and eventually just
had enough customers who hadleft Heroku to AWS and wanted to
use JudoSkill.
I'm like this is stupid, thatwe're not taking the time to
investigate this and buildsomething.
So, yeah, that was morechallenging because I didn't
really have any experience withAWS beyond, like creating an S3
bucket.
That's the extent of what I haddone in AWS at that point.
(10:48):
So, yeah, kind of had to learnhow ECS worked and the
difference between using it withEC2 and Fargate, all the
acronym soup that's in AWS.
Oh man, the permissions withIAM, all that stuff.
That was a painful integrationfor sure, and I still dread
anytime I have to get into AWSfor anything.
Speaker 1 (11:10):
Is it to the point where you're like, okay, that's
done, I can mostly just let itsit there.
Or is it?
No, there's still a bunch ofstuff to do, but it's good
enough for now.
I can just touch it when I'mfeeling.
Pain is going to be fun today,I don't know.
Speaker 2 (11:22):
We've got a list of quality of life improvements we
want to make on that platformand, more generically, on sort
of all the non-Horoku platformsthat we've integrated with.
It just works and we don'treally have to dig into it for
the most part Just a lot oflittle improvements that we want
to make.
And then the biggest changethat is on our radar that I
either want to make this monthor next month is redoing our
(11:46):
pricing for the non-Horokuplatforms On Heroku.
Because we're in themarketplace, we're pretty much
tied down.
We have to do monthly tieredfixed price billing and when we
did our rendering, ECSintegration made the decision to
go with metered billing basedon how much compute usage only
because I couldn't think of agood way to do the flat tiered
(12:08):
billing because of the way ECSand render work.
But it's been a pain.
Meter billing is such a pain Iwant to throw it out and we're
going to throw it out.
We're going to migrate to justflat tiered pricing.
It'll make our life easier.
It'll make our customers' liveseasier.
It'll make it easier tounderstand what your price is
going to be before you sign upfor it Another one of those
(12:29):
decisions that I'm having toundo from the initial render and
ECS integration I feel yourpain.
Speaker 1 (12:36):
Billing anytime is scary and difficult Figuring out
how to do billing and then,even whenever you implement
anything, any change feelsdangerous because you're dealing
with people's money.
It's terrifying.
Speaker 2 (12:50):
Yeah, that was one of the best things about building
on Heroku is that we didn't haveto think about billing.
Heroku handles all the billing.
They charge a premium for it.
It's an app store like Applewhere they take a 30% cut.
But the reality is I wouldn'tbe doing this full time if it
weren't for Heroku's add-onmarketplace, because I was able
to just throw it in there and,with almost no marketing, people
(13:12):
would just find it and sign upfor it and start using it.
So it just kind of grew slowlyon its own.
So I've never really had aproblem paying that tax to be in
the Heroku marketplace.
But yeah, when we expandedbeyond that, yeah, we had to
handle the billing and so thatwas a whole new world and it is
super stressful.
Speaker 1 (13:31):
So, now that you have to have billing outside of the
heroku marketplace, is the ideato?
Maybe it'll still be in theheroku marketplace?
Because why not?
But also we would prefer if youdidn't for heroku customers.
Speaker 2 (13:44):
We're definitely still going to be in the
marketplace and we'll stillencourage that they go through
the marketplace.
But, yeah, for anybody who'snot on heroku, obviously, when
it comes to which hostingplatform to use, I'm still a fan
of Heroku.
I still think Heroku is a greatchoice and we're still hosted
on Heroku mostly.
Yeah, we actually just movedour staging environment to
(14:06):
render a month ago, but we'reabout to move it back.
That's a whole nother thing.
But, yeah, definitely still afan of Heroku.
Speaker 1 (14:12):
If anybody's using Heroku, we're going to encourage
them to go through themarketplace.
We're not going to not diveinto staging on render, but not
for much longer.
Was that an experiment?
Was that just hey, let's seehow render works.
What was that move for and whyare you moving back?
Speaker 2 (14:28):
It was sort of all of the above.
It was my own curiosity.
I wanted to see what's it likerunning an actual product on
render.
I wanted to see, cost-wise, howis that going to compare?
Because our staging environmentwe call it a staging
environment, but it's a veryproduction-like environment Our
staging app actually auto-scalesour production app and our
(14:48):
production app auto-scales ourstaging app, so our staging app
still gets pretty decent traffic.
So I wanted to see how the costwould compare.
I wanted to see how theexperience would compare and I'm
trying to do more video contentand that kind of thing and I
thought this would be a goodopportunity for some video
content to show what's it liketo migrate an app from Heroku to
Render and then do somecomparison there.
(15:09):
But yeah, what's beeninteresting?
After leaving it there for amonth, I actually just got our
first full month bill for theRender app and it's crazy,
looking at the compute a lotcheaper than Heroku.
But what was surprising isRender charges you for egress,
(15:30):
so the outgoing bandwidth thatHeroku doesn't charge you for,
and that was 80% of our bill onrender and that's something we
didn't really account for whenwe experimented with the move
and I'm like, yeah, if we moveour product to render, it's
going to be more expensive thanHeroku.
So that, amongst a few other,just little bumps along the road
(15:50):
that just make Heroku asmoother experience.
Speaker 1 (15:55):
And I mean, if you're more comfortable and familiar
with it, why not right?
If it's not saving money orsaving headache or anything, why
bother?
Speaker 2 (16:03):
Yeah, but I definitely want to be able to
serve our customers better.
I want to be even more familiarand comfortable with all the
platforms that we support, soI'm going to experiment with
moving to Railway and fly, justto see what it's like.
Speaker 1 (16:16):
Yeah, it makes sense.
Hey, these are the products weoffer and here's if you're
trying to move between them oranything, having that video
content is always a nice thing.
Speaker 2 (16:25):
I appreciate video content like that we got a lot
of people asking us for adviceon how to choose between them,
and it's hard because at the endof the day they all do the job
well enough and it's kind oflike try them out and see which
one feels the best.
It's a lot of work to try themout all individually, so maybe
(16:47):
if I can try them all out andproduce enough content to help
out with that, yeah, even whatyou learned with render is oh
yeah, the compute's about thesame, so that's about the same.
Speaker 1 (16:56):
But if you have a high Ingress app, that's where
Render gets expensive and maybeother people wouldn't know that.
And that's a what is your app,what is your app doing, kind of
follow-up question where westart getting into.
Maybe there is one that isbetter than the other for
certain types of apps.
Speaker 2 (17:14):
Yeah, one of the things that I want to build in
the next month or two is apricing calculator that covers
all of the popular pass servicesHeroku, render, fly, railway
because they all charge sodifferently.
They all have their own pricingcalculators that work
differently.
Some of them charge for egress,some of them don't.
Some of them charge for teammembers, some of them don't.
(17:35):
Some of them charge for teammembers, some of them don't.
Especially as somebody who isstill a fan of Heroku and sees a
lot of the hate that Herokugets on social media and whatnot
, it's not as clear cut as youthink.
Yeah, heroku does have somedynos per fem that are
ridiculously expensive and youshould basically never use, but
(17:55):
it doesn't mean that it's alwaysgoing to be astronomically more
expensive than the alternatives.
Speaker 1 (18:00):
Hidden costs and fees , and Ticketmaster springs to
mind.
Speaker 2 (18:05):
Yeah, I'm not going to label any of them with
Ticketmaster.
That's brutal.
Speaker 1 (18:09):
So what kind of blockers do you currently have?
Or, if there isn't a blockerthat you're currently dealing
with, what's a recent one thatyou had and what did you do to
get yourself around it or over?
Speaker 2 (18:19):
it need to put some time and energy into marketing.
So just making that time formarketing and forcing myself to
do the hard work of marketing,that's just been a continual
(18:41):
challenge.
This year I hired my firstfull-time developer, carlos
Antonio, who's a Rails corecontributor, and I worked with
him at a previous job.
He's working with me full-timenow.
So that has helped me put somemore time into marketing because
I know he is going to keepmoving the product forward even
when I'm not Even with that.
(19:04):
It's hard, because I reallyreally like working on the
product.
There's a lot of things that Iwant to get done.
One thing that I'm startingjust this week actually, is a
couple of friends of mine,colleen Schnettler and Leanna
Patch.
They are starting a new thingthey're calling SaaS Marketing
Gym, where I think it's mostlykind of solo founders like me or
(19:26):
similar get together for anhour a day.
It's like going to the gym.
It's like an hour set asidewhere we're coming together, we
have our goals in marketing andwe're just going to focus on our
marketing tasks and Leanna andColleen are going to be there to
help us get on, stock andbounce ideas off of and stuff
like that.
So actually doing my first gymsession right after we finished
(19:47):
recording this podcast.
Yeah, I'm super excited aboutthat.
I think that's going to help mekeep moving the ball forward on
the marketing side of things.
Speaker 1 (19:56):
Yeah, that's interesting.
I've talked to a few people nowrunning their own SaaS or
either as a solo or co-foundersetup, but that seems to be the
blocker.
I don't know how to market.
I don't know how I'm supposedto drive up people knowing about
my product and without justtalking about it 24-7, which
(20:19):
feels weird.
If I ever came up with a goodenough idea to run with, that
would be my big blocker too.
I don't know how to tell peopleabout my thing that I'm doing,
and I wouldn't even know whereto go to learn.
You want to go learn how tocode?
There's a number of differentplaces that will teach you how
to code.
Do those places exist formarketing, like a marketing
(20:39):
bootcamp?
Here's how to talk to people.
Here's why this phrasing worksbetter than this phrasing.
I run into it with designers.
Bump up that line by one pixeland use this font instead.
The entire page changes Likehow the heck did you see that?
Speaker 2 (20:54):
I can't see that until it's done, that is a big
reason that I joined TinySeed isto get community and mentorship
around marketing, because I didfeel that was my huge gap, and
part of it is knowing what to do, and the other part is just
making yourself do it.
Because, as a developer and Iknow a lot of developers can
sympathize with us I don't likebeing marketed to, so I don't
(21:17):
want to be the person who'salways talking about my product,
who's always pushing it onpeople.
Obviously, there are tastefulways to do it, there are
authentic ways to do it, and sothat's the challenge is how to
promote your product, buildawareness around your product,
but in an authentic way thatyou're genuinely trying to help
people and contribute to theoverall community, rather than
(21:41):
just tell people to use yourthing.
Speaker 1 (21:43):
Well, I think your first judo scale commercial was
a huge hit.
I loved it at the very least.
I got to ask how did you comeup with that?
Speaker 2 (21:53):
You tweeted that you were looking for companies to
sponsor prizes for the game showat RubyConf and I reached out
to you.
I was like, yeah, we'd love tosponsor some prizes.
And that's when you threw outthat, yeah, I've got extra time
this year, so if you want torecord a 30 second commercial,
I'll show it on stage.
The only requirement is that ithas to be funny.
And I was just oh well, I'm notgoing to do that.
(22:15):
That's way too much pressure.
I don't consider myself a funnyperson Like.
That's not me.
I was talking to a buddy of minewho actually helps me out on
Judo scale.
So, in addition to Carlos,who's full-time on Judo scale a
friend of mine, local to centralOhio where I live, he spends a
few hours a week with me on judoschool.
(22:35):
We get together at a coworkingspace and just kind of pair on
things and bounce ideas off ofeach other.
I was talking to him about it.
One of the reasons I loveworking with him is he always
pushes me out of my comfort zone.
He always challenges me, and soI told him about this.
This would be a hugedistraction.
I don't even know where we'dbegin.
He's like oh, we're definitelydoing it.
We're definitely doing it, yes.
(22:55):
So yeah, he and I just went outto lunch and started throwing
some ideas around.
Yeah, he was the one who came upwith the Billy Mays idea and
then we actually used ChatGPT towrite the initial script for it
.
We had the idea of let's havean actual scale.
The actual scale is the judoscale and that's what we're
(23:19):
selling in this infomercial.
So we kind of described that toChatGPT and it wrote the
initial script, nailed theinfomercial Billy Mays tone and
everything.
We tweaked it a ton but itreally helped get us past the
initial blank page.
Where do we start?
It gave us a starting point.
We iterated on that.
Once we felt good with thescript, timed it and got it to
(23:41):
that 30 second mark, we gottogether in his kitchen and
recorded it in a couple hoursand I spent maybe a day editing
it.
Probably didn't even need tospend that much.
I was just having way too muchfun adding sound effects and
making the cuts and everything.
It was so much fun.
Speaker 1 (23:58):
It turned out so great.
I had that idea when I was atRailsworld I was talking to Josh
Wood about a new Honeybadgerwas going to be sponsoring,
because they sponsor the podcastnow and I just, man, when I
first first did this whole gameshow thing, I thought about
having a commercial.
When I first first did thiswhole game show thing, I thought
about having like commercial.
I don't know if you've everseen any of wise old conference
(24:22):
talks, but he was just super outthere and he would throw random
commercials in his talks.
It wasn't a commercial, but itwas like a video that came
across like a commercial, but itwould talk about Ruby stuff.
Man, it'd be so funny to docommercials or something, and
he's.
I think we could do commercials.
Why not?
If you've got the hour now fora keynote, you could if you
wanted to, and I was all right.
Well, if you give me a 20 to 30second commercial that's funny
Then I'll roll with it and I'lltalk to anyone else, offer it up
(24:45):
, and I sent it to you too, andthis is great way of being like
thank you for sponsoring theprizes and also adding to the
game show atmosphere.
And, man, when I saw yours, Idon't know if I could even play
this on stage, because I can'tstop laughing.
It was so good, it's so good.
I'm going to link it when Ipost about this.
(25:07):
I think it's on your Blue Skyaccount too.
You posted the full video, butI'll make sure it's in my post
about this episode, because itis just so, so good.
Speaker 2 (25:18):
Thank you, I appreciate it.
We got a lot of good feedbackat the conference.
The commercial didn't evenreally explain what Judo Scale
does.
Speaker 1 (25:25):
Whatever, it's fun.
It got people to Google JudoScale, I guess, so okay.
So the last question other thanyour new career in infomercial
production, what is somethingcool, new or interesting that
you've recently learned,discovered?
Got into hobby-wise, code-wise,doesn't matter.
What is something you want toshare?
Speaker 2 (25:47):
It actually is related to the commercial,
because I have been loving usingDescript for making video
content.
I've been trying to make somemore video content, whether it's
for Blue Sky or LinkedIn orYouTube or a RubyConf commercial
, and if you want to get intovideo content at all, I can't
think of any reason not to useDescript.
(26:09):
I'm not getting paid by them,it's just phenomenal.
So Descript is a video editingapp.
You record in it, ittranscribes your video and then
you edit the script of the videoand that edits the video for
you.
You cut out words from thescript.
It cuts out that part of thevideo.
It actually does a really goodjob with the edit points too.
Sometimes I have to tweak them,but most of the time I don't.
(26:31):
Really good job with the editpoints too.
Sometimes I have to tweak them,but most of the time I don't.
It can do the captions, it cando highlight reels.
It can do so much.
So yeah, if you want to getinto video content, you've got
to give Descript a try.
Speaker 1 (26:41):
It makes it so easy?
Yeah, and I'll expand on that.
If you are a new podcaster andyou are not ready to have your
own editor, but you want toclean up the audio feed, I used
the script in the beginning too.
It's really easy to reducepauses.
It'll automatically take outuhs and ums and all those filler
words.
Yeah, it also could edit thescript right there.
So you're like oh, I have toremember to take out that
(27:03):
bumbling run-on sentence that Ihad there.
You could just delete it and itwill delete it for you.
Speaker 2 (27:08):
You might tweak the audio a little bit, but yeah
it's a very cool application andit even has a library of sound
effects and things like thesound effects that I put in the
commercial.
Those were in Descript's soundeffects library that I was just
able to throw in there.
It was awesome.
Speaker 1 (27:23):
That's a good one.
Yeah, I think that Descript isa wonderful product and has a
lot of uses Video.
I've used it for internalvideos when I was at within
three and, like we would gothrough refactoring, I would
just record it.
It was so cool because I don'tknow if I can do a video this is
going to be crazy.
But I was using Descript to dopodcasting already and I was all
right well, try their videofeature and it made it so dummy
(27:45):
easy to do what I was trying todo to do.
I was very impressed by thefinal product to the point where
I was wow, maybe this is athing, and then it's not for me,
but it's way too much work andI'm entirely too much of a
perfectionist to do that.
But yeah, it's such a cool tool.
I believe my podcast editor,paul, who does a bunch of the
other podcasts, edits inDescript also, so it is
(28:09):
definitely a tool for theprofessionals too.
Speaker 2 (28:11):
Yeah, and the other thing I would add is, if anybody
is kind of curious about doingvideo content but is intimidated
by it at all, especially ifyou're thinking about needing to
get specialized equipment orwhatever.
The other thing that I wouldrecommend, if you have a modern
iPhone and a modern Mac, youriPhone is an excellent webcam.
I don't know if you publishedthe video for these things.
(28:31):
The camera that I'm using rightnow is my iPhone, just on a
magnet clip at the top of mymonitor, and, yeah, continuity
mode is built into the newerMacs and I think it looks great.
Obviously, you can get DSLR orwhatever and get marginally
better video, but I think thestep up from a traditional
webcam to using your iPhone as awebcam is huge and if you
(28:54):
already have those things, youcan make pretty good videos with
minimal equipment.
You got to get a good mic and amic arm.
That's the biggest thing.
And, yeah, then use your iPhone, use Descript, you can make
good stuff.
Speaker 1 (29:06):
I don't publish the videos.
I've thought about it whenYouTube started hosting podcasts
, but no one wants to see thisface.
I don't want to see this facethat much.
So I try not to look in themirror too often, even, but to
piggyback on the.
You can make minor improvements.
Even lighting makes a hugedifference.
I have some cool lights.
I can turn them off.
Video is still fine, but to meit gets much better once you
(29:31):
introduce some really goodlighting too.
So for sure, you know, itdoesn't have to be the most
spectacular camera I definitelydon't have a very fancy camera
but the lighting makes a hugedifference.
I even run the lighting justwhen I'm on a zoom call, just
because it looks marginallybetter.
So why not?
So yeah, you can use youriphone, you can use some extra
(29:52):
lighting.
Doesn't have to be anythingfancy, just a lamp with some
more light will do.
But yeah, I think themicrophone, the audio, is in
many ways the more important bitof even a video, because you
got to hear what's going on.
So is there anything else thatyou wanted to talk about?
Touch on, discuss before wewrap.
Speaker 2 (30:13):
I'm talking about RailsConf.
How are you feeling aboutRailsConf in Philadelphia?
Does that excite you at all?
I'm so excited.
Speaker 1 (30:19):
Could you tell at RubyConf me bouncing around like
an over-caffeinated squirrel.
I am pumped.
We've never had a Ruby orRailsConf in Philadelphia, which
is horrible in my opinion.
Railsconf in Philadelphia,which is horrible in my opinion.
But that will all change inJuly when RailsConf comes to
Philly.
The very last RailsConf will bein Philly.
(30:39):
I'm super pumped.
I'm actually going down to thecity this weekend for a
different event, but it'll be inthe same area of operation as
where the conference will be.
So I'm going to do a little bitmore than just my mindless
walking around and just, oh,that smells good.
I'll go in this restaurant.
I'm going to kind of hit someplaces that I know I want to get
good information on and sendthat back to Ruby Central to try
(31:01):
and help make a visitor's guideand the big things like
Independence Hall and theLiberty Bell People know about
that.
But what else is there in Philly?
What's surrounding Philly?
In case you're coming out for alonger period of time and
you're bringing the familybecause it's in July instead of
a weird in school month thistime there's a lot of cool stuff
around Philly, in Philly.
So I'm very excited to do thatkind of stuff, see how many
(31:23):
people come out and go.
Yeah, you're right, philly isreally cool.
I'm glad we had this experiencehere.
But, yeah, great food.
The area that the conferencewill be in is a lot of fun, so
very, very excited.
Speaker 2 (31:35):
I'm excited too.
It's going to be the firstconference that we'll be an
official sponsor of.
We're going to have a booth andhoping to get the whole
JudoSkill team out there allthree of us.
We're talking about beinguncomfortable doing the
marketing stuff.
The idea of standing at a boothand having to talk to people
about JudoSkill as they walk upto it is absolutely terrifying
to me.
It's time.
It's time to force myself outof my comfort zone and do that.
Speaker 1 (31:59):
You can also take a page out of Honey Badger's book.
I think Ben and Josh do a greatjob.
They just have this booth withcouches and they're there and
they're chilling and maybethey're hacking on something
with someone and they've got abanner and if someone comes up
and they go, hey, what's HoneyBadger?
They'll get up and they'll talkto them, but otherwise it's
just a cool space to hang outand chill and I feel like
everybody knows what HoneyBadger is.
Oh, definitely that kind ofpassive marketing in a way at a
(32:25):
place like RailsConf can beawesome and you'll have a TV
playing your Billy Mayscommercial.
Speaker 2 (32:31):
There's only so many honey badger ideas that I can
steal before it just becomesdisgusting.
So I love their lounge wherecome, hang out and chill.
We'll have to think ofsomething unique to do with our
booth.
Speaker 1 (32:38):
What's the swag situation going to be like?
What are you doing?
Don't know.
Yet I start brainstorming.
That's the key.
That's how you get them.
When you give them cool stuff,they'll remember you forever.
Speaker 2 (32:47):
My one rule is whatever swag we bring, it's
going to be a one size fits all.
We're not bringing differentsizes of t-shirts.
Figure, how many of each sizedo you need to bring?
Like no, it's just going to bewhatever, it is One size fits
all.
Speaker 1 (33:02):
Don't want an extra suitcase like Josh brings of
just a ton of t-shirts in agiant rolling case, because I
really bring a million of them,Although I will say you do have
very cool t-shirts.
I love the auto scale and chillt-shirt.
They're very soft andcomfortable and they come in a
variety of colors.
But the auto scale and chill isfunny to me.
It makes sense and it's a verygood t-shirt.
Speaker 2 (33:24):
I did an initial run of rails auto scale t-shirts
back in the day, had them madein a bunch of different sizes
and stuff like that, and then itended up just with a closet
full of t-shirts.
And then we rebranded and stillhave a closet full of t-shirts.
I decided I wasn't going tomake that choice again.
So, yeah, when we made the judoscale shirts, we're going to do
them print on demand as peoplewant one, which, yeah, if you're
(33:44):
listening to this and you wanta judo scale shirt, shoot me an
email, adam, at judoscalecom.
Speaker 1 (33:49):
I'll get you a t-shirt.
I recommend them.
They're comfy and hilarious.
So I was wearing mine atRubyConf and I will be wearing
mine at RailsConf.
Well, cool man, how can peoplefind you on the internet?
Yeah, I'm.
Speaker 2 (34:01):
AdamLogic on BlueSky.
Adamlogic on LinkedIn.
We are judoscalecom.
If you're interested inauto-scaling your web app, yeah,
I think that's it Wherever andauto-scaling your web app yeah,
I think that's it.
Speaker 1 (34:11):
Wherever it may be hosted.
Adam at JudoScalecom.
Speaker 2 (34:13):
I'm not covering that hosting platform.
I'm still old school and loveemail.
Shoot me an email,adamatjudoscalecom.
I'd love to hear from you.
Awesome, awesome.
Speaker 1 (34:20):
Billy.
Speaker 2 (34:21):
Oh yeah.
Speaker 1 (34:22):
So excited Yay to chat with us.
Thank you.
Speaker 2 (34:25):
Drew, this was a lot of.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.