libaws VS Concourse

i print the error along with file and line number every time i return it. clunky, but it works.

in fact i print file and line with every log message.

https://github.com/nathants/libaws/blob/87fb45b4cae20abd1bb1...

cloud is so good now it’s hard to justify not doing something bespoke. ec2 spot is insanely cheaper than turnkey cicd, and better in almost every way.

i’m delighted to pay 30% over infra cost for convenience, but not 500%. and it better actually be convenient, not just have a good landing page and sales team.

this month i learned localzones have even better spot prices. losangeles-1 is half the spot price of us-west-2.

for a runner, do something like this, but react to an http call instead of a s3 put[1].

for a web ui do something like this[2].

s3, lambda, and ec2 spot are a perfect fit for cicd and a lot more.

1. https://github.com/nathants/libaws/tree/91b1c27fc947e067ed46...

2. https://github.com/nathants/aws-exec/tree/e68769126b5aae0e35...

like linux, cloud is a lot to learn, but worth it.

like linux, cloud is best kept simple, or it can become brittle and confusing.

like linux, cloud has a lot of cool things like zfs, that should be appreciated but rarely used.

like linux, using go makes your life a lot easier. the aws go sdk is the documentation.

like linux, you have to learn a lot and then find the core utility you actually care about. for me it is:

https://github.com/nathants/libaws

the good new is, for the 95% of projects that can tolerate it, aws the good parts are actually both simple and easy[1].

it’s hard to find things you can’t build on s3, dynamo, lambda, and ec2.

if either compliance or a 5% project demand it, complicated solutions should be explored.

1. https://github.com/nathants/libaws

i also wanted a good cli for aws, and built one:

https://github.com/nathants/libaws

companies like fly are fantastic.

they provide a good service, and they put market pressure on aws.

cool transition and fun writeup!

for low, intermittent traffic sites, go on lambda might be a better comparison:

https://github.com/nathants/libaws/tree/master/examples/simp...

lambda + s3. add ec2 spot if you need it.

just make sure you understand how billing works. mostly it’s just egress bandwidth is expensive.

do something like this:

https://github.com/nathants/aws-gocljs

or with less opinions:

https://github.com/nathants/libaws/tree/master/examples/simp...

welcome to cloud, glhf!

aws ux for retaining both hair and sanity.

https://github.com/nathants/libaws

build and scale systems with artificial load on aws! scaling the load testing will be just as interesting as scaling the system under test.

start with low bottlenecks, ie a cluster of c6i.large ec2 spot. how fast can you do this? have fast can you scale that? ec2 and s3 is all you need to build anything.

use ec2 spot, avoid network egress, avoid cross region/zone traffic, create and destroy ec2 instances as needed instead of letting them sit idle. you could grow system scaling intution for the price of your streaming subscriptions.

start with something like this:

https://github.com/nathants/libaws/tree/master/examples/comp...

maybe mess around with public datasets on aws, just make sure to be in the correct region to avoid data egress.

welcome systems friend. one accurate measurement is worth a thousand expert opinions. scaling is fun!

aws has too many knobs, presumably to satisfy the union of the needs of all the enterprise customers. that said, lambda+s3+dynamodb+ec2 are pretty good once you tape over all the knobs that aren't needed. i work with them like this[1].

these days i build on aws and r2. aws for the nuts and bolts, r2 for high bandwidth egress. it's a perfect match.

1. https://github.com/nathants/libaws

Ableton ⬩ Acima ⬩ ACKO ⬩ ActiveState ⬩ Adrima ⬩ AJR International ⬩ Alma ⬩ Astrosat ⬩ Ava ⬩ Avetta ⬩ Azara ⬩ Barmenia ⬩ Basiq ⬩ Beautiful Destinations ⬩ BEC Systems ⬩ Bekk ⬩ Bellroy ⬩ Bendyworks ⬩ Bernoulli Finance ⬩ Blue Fog Training ⬩ BravoTran ⬩ Brilliant ⬩ Budapest School ⬩ Buildr ⬩ Cachix ⬩ CalculoJuridico ⬩ CareRev ⬩ CARFAX ⬩ Caribou ⬩ carwow ⬩ CBANC ⬩ CircuitHub ⬩ CN Group CZ ⬩ CoinTracking ⬩ Concourse CI ⬩ Consensys ⬩ Cornell Tech ⬩ Corvus ⬩ Crowdstrike ⬩ Culture Amp ⬩ Day One ⬩ Deepgram ⬩ diesdas.digital ⬩ Dividat ⬩ Driebit ⬩ Drip ⬩ Emirates ⬩ eSpark ⬩ EXR ⬩ Featurespace ⬩ Field 33 ⬩ Fission ⬩ Flint ⬩ Folq ⬩ Ford ⬩ Forsikring ⬩ Foxhound Systems ⬩ Futurice ⬩ FörsäkringsGirot ⬩ Generative ⬩ Genesys ⬩ Geora ⬩ Gizra ⬩ GWI ⬩ HAMBS ⬩ Hatch ⬩ Hearken ⬩ hello RSE ⬩ HubTran ⬩ IBM ⬩ Idein ⬩ Illuminate ⬩ Improbable ⬩ Innovation through understanding ⬩ Insurello ⬩ iwantmyname ⬩ jambit ⬩ Jobvite ⬩ KOVnet ⬩ Kulkul ⬩ Logistically ⬩ Luko ⬩ Metronome Growth Systems ⬩ Microsoft ⬩ MidwayUSA ⬩ Mimo ⬩ Mind Gym ⬩ MindGym ⬩ Next DLP ⬩ NLX ⬩ Nomalab ⬩ Nomi ⬩ NoRedInk ⬩ Novabench ⬩ NZ Herald ⬩ Permutive ⬩ Phrase ⬩ PINATA ⬩ PinMeTo ⬩ Pivotal Tracker ⬩ PowerReviews ⬩ Practle ⬩ Prima ⬩ Rakuten ⬩ Roompact ⬩ SAVR ⬩ Scoville ⬩ Scrive ⬩ Scrivito ⬩ Serenytics ⬩ Smallbrooks ⬩ Snapview ⬩ SoPost ⬩ Splink ⬩ Spottt ⬩ Stax ⬩ Stowga ⬩ StructionSite ⬩ Studyplus For School ⬩ Symbaloo ⬩ Talend ⬩ Tallink & Silja Line ⬩ Test Double ⬩ thoughtbot ⬩ Travel Perk ⬩ TruQu ⬩ TWave ⬩ Tyler ⬩ Uncover ⬩ Unison ⬩ Veeva ⬩ Vendr ⬩ Verity ⬩ Vnator ⬩ Vy ⬩ W&W Interaction Solutions ⬩ Watermark ⬩ Webbhuset ⬩ Wejoinin ⬩ Zalora ⬩ ZEIT.IO ⬩ Zettle

> Why bother, when Dagger caches everything automatically?

The fear with needing to run `npm ci` (or better, `pnpm install`) before running dagger is on the amount of time required to get this step to run. Sure, in the early days, trying out toy examples, when the only dependencies are from dagger upstream, very little time at all. But what happens when I start pulling more and more dependencies from the Node ecosystem to build the Dagger pipeline? Your documentation includes examples like pulling in `@google-cloud/run` as a dependency: https://docs.dagger.io/620941/github-google-cloud#step-3-cre... and similar for Azure: https://docs.dagger.io/620301/azure-pipelines-container-inst... . The more dependencies brought in - the longer `npm ci` is going to take on GitHub Actions. And it's pretty predictable that, in a complicated pipeline, the list of dependencies is going to get pretty big - at least a dependency per infrastructure provider we use, plus inevitably all the random Node dependencies that work their way into any Node project, like eslint, dotenv, prettier, testing dependencies... I think I have a reasonable fear that `npm ci` just for the Dagger pipeline will hit multiple minutes, and then developers who expect linting and similar short-run jobs to finish within 30 seconds are going to wonder why they're dealing with this overhead.

It's worth noting that one of Concourse's problems was, even with webhooks setup for GitHub to notify Concourse to begin a build, Concourse's design required it to dump the contents of the webhook and query the GitHub API for the same information (whether there were new commits) before starting a pipeline and cloning the repository (see: https://github.com/concourse/concourse/issues/2240 ). And that was for a CI/CD system where, for all YAML's faults, for sure one of its strengths is that it doesn't require running `npm ci`, with all its associated slowness. So please take it on faith that, if even a relatively small source of latency like that was felt in Concourse, for sure the latency from running `npm ci` will be felt, and Dagger's users (DevOps) will be put in an uncomfortable place where they need to defend the choice of Dagger from their users (developers) who go home and build a toy example on AlternateCI which runs what they need much faster.

> I will concede that Dagger’s clustering capabilities are not great yet

Herein my argument. It's not that I'm not convinced that building pipelines in a general-purpose programming language is a better approach compared to YAML, it's that building pipelines is tightly coupled with the infrastructure that runs the pipelines. One aspect of that is scaling up compute to meet the requirements dictated by the pipeline. But another aspect is that `npm ci` should not be run before submitting the pipeline code to Dagger, but after submitting the pipeline code to Dagger. Dagger should be responsible for running `npm ci`, just like Concourse was responsible for doing all the interpolation of the `((var))` syntax (i.e. you didn't need to run some kind of templating before submitting the YAML to Concourse). If Dagger is responsible for running `npm ci` (really, `pnpm install`), then it can maintain its own local pnpm store / pipeline dependency caching, which would be much faster, and overcome any shortcomings in the caching system of GitHub Actions or whatever else is triggering it.

> Imagine you live in a world where no part of the build has to repeat unless the changes actually impacted it. A world in which all builds happened with automatic parallelism. A world in which you could reproduce very reliably any part of the build on your laptop.

That sounds similar to https://concourse-ci.org/

I quite like it, but it never seemed to gain traction outside of Cloud Foundry.

I used Concourse[0] for a while. No real complaints, the visibility is nice but the functionality isn't anything new.

[0] https://concourse-ci.org/

We run https://concourse-ci.org/ on our own hardware at our office. (as a side note, running your own hardware, you realise just how abysmally slow most cloud servers are.)

Concourse: https://concourse-ci.org

We use https://concourse-ci.org/ at the moment and have been reasonably happy with it, however it only has support for linux containers at the moment, no windows containers. (MacOS doesn't have a containers primitive yet unfortunately)