alexellisuk

Spoiler: I'm the OpenFaaS Founder, so am biased here. This is probably our best tutorial so far on automating OpenFaaS installation and function updates using Argo CD. We've had previous articles, but I'd say this is the most concise and also the only one to cover the Image Updater. Argo's Image Updater watches a container registry and updates your Helm chart's YAML files automatically.

You can apply what we've covered for your own CRDs or Kubernetes objects, so feel free to take a peak even if you're not an OpenFaaS user yet.

This article is from the OpenFaaS blog, spoiler: I'm the OpenFaaS founder. We've found for a number of users, especially large commercial ones, that building programmatically is important.

"faas-cli build" is still the bread and butter for free and community users, but we wanted to show what's possible. Hope you find this interesting.

I'm posting this from the OpenFaaS blog - it's based upon discussions / patterns we've seen with customers adopting CI/CD for functions.

It could also be useful for other CRDs that you're creating or managing.

It's Han's second post on the blog.

I'm the author of inlets - this article was put together by an inlets/openfaas community member. I thought it might be useful to share.

For anyone who is curious, this already includes the container image on the Docker Hub.

https://twitter.com/n_copa/status/1528822310957891584?s=20&t...

_Bring_ _back_ _Geocities_

> There are a number of tools that can be used to manage a fleet of hosts, however they each involve significant redesign of your system. Not all of them are going to slot easily into an existing estate of infrastructure tooling.

> After a recent call with a customer, who is using inlets tunnels to manage remote appliances for customers, I wanted to show you what this looks like, and a sample architecture with Kubernetes.

This is not my book, I'm not an affiliate either, but I follow Philippe and also have authored a book on Go on Gumroad, so wanted to support him. The price seems very reasonable for an exploratory topic IMHO.

I can't help think that these mandates are being made by people who are out of touch with everyday people.

Identifying an unmet need, getting a market fit for a solution, understanding what people are willing to pay are all counter-intuitive for engineers, but essential to make money.

These are my recommendations to anyone who's here and not getting the traction they want, or are considering starting a business. The 1-Page Marketing Plan - Alan Dibb. The Right It - Alberto Savoia. Monetizing Innovation - Georg Tacke & Madhavan Ramanujam. Obviously Awesome - April Dunford

You can get bare-metal instances on AWS :) I've added a link to another repo I maintain that lists all the places I know of to get bare-metal servers. Hope that helps.

Hey, I'm the maintainer of OpenFaaS. Thanks for mentioning the project.

Pipedream v1 ran on Lambda, but this post is touted as "AWS Lambda alternative", have you moved off the AWS Lambda platform?

I was curious how you added support for other languages, was it in the same way that you built-in support for Node.js? Where all the code gets effectively built into a single Lambda function and proxied?

For people who want to call OpenFaaS workloads from Pipedream, that's possible over HTTP and Pipedream makes that relatively simple. One thing I think they've really nailed is the OAuth integration, it just makes that whole painful flow a lot more convenient. I also like the editor and the way it infers what kind of input will go into the next step in a workflow.

Alex

Ignite will launch a microVM in a similar way to my lab, except it has a user experience designed more around Docker's CLI. Great if you just want to use it like a black box.

If you're more interested in integrating with Firecracker into your own systems - come to the session because I'm covering this in detail.

Weaveworks is now working on a new tool to create K8s clusters where each node is a microVM - Richard Case will be doing a demo of that, and it's super fast. It's for a Telco customer they have, so it's a real use-case, not just shiny tech, so should score extra points, right?

It depends. In a container, you can't really run docker itself or full K3s for instance.

So if you have workflows which that you're struggling to make work in a container, without root, a microVM gives you a full environment, where you can do whatever you want.

Of course, we are covering all this in the live-stream, see the link at the end of the repo!

Many of the examples you'll find are broken due to changes in Firecracker 1.0 - the official quickstart guide doesn't cover the most interesting thing - working Internet access - or extracting a filesystem from a container.

Edit: I've added a link for a live-stream I'll be doing with Richard Case from Weaveworks on use-cases, differences vs. containers and legacy VMs - and a number of demos.

> Many of us own at least one Raspberry Pi, and if it's not doing duty as a media player, retro gaming console, blocking ads, or reporting the weather, then it may well be gathering dust.

> I'm writing this article as an excuse for you to blow the dust off those microchips, and to put your pocket-sized silicon to work as the world's smallest API-driven cloud.

> By following the instructions in this article, you'll be able to deploy scheduled tasks, webhook receivers, web pages and functions to your pocket-sized cloud from anywhere using a REST API.

I just published: A case-study in end-to-end testing with Go - using a recent challenge to show you some tools and techniques that you could use in your own work.

What other techniques do you use for your end to end testing? How do you keep them from being too brittle and time-consuming? What other kinds of testing do you do?

If you want to reach out to me with questions, feel free to ping me over on Twitter - https://twitter.com/alexellisuk

Full disclosure - I'm the author of the open-source OpenFaaS project, and wrote this blog post.

The aim is to show existing openfaas users how to use React, and to show how users deploying React to a CDN can use functions for dynamic data through a CORS policy.

OpenFaaS has a predictable, low cost since it's self-hosted. Feel free to see the post for more on that.

Second time of posting to HN.

Thanks for the comment John. You might want to take a look at "faasd" that I mention, the images once built can deploy there or to an existing K8s cluster with the openfaas helm chart.

This is a guest post by Nitish. He shares his experiences from using OpenFaaS at work, to becoming a community contributor.

> "In this short post, I’m going to tell you a bit about how I found out about openfaas, what my team did with it at work, and then how I’ve been involved as a contributor. Finally, I’ll give you some tips for getting started. Contributing to open source can be really rewarding and it also helped with my own career development."

Feel free to ask questions over on Twitter (@alexellisuk) - we're also welcoming to new contributors who want to get into Go and community projects. The focus is on developer tooling and cloud native. Let us know if that's of interest.

Full disclosure - I'm the author of the open-source OpenFaaS project, and wrote this blog post.

The aim is to show existing openfaas users how to use React, and to show how users deploying React to a CDN can use functions for dynamic data through a CORS policy.

OpenFaaS has a predictable, low cost since it's self-hosted. Feel free to see the post for more on that.

Do you also act as an SRE for those apps hosted on your bespoke hosting platform?

How does "just using ingress" work on your local cluster with K3d/KinD/minikube/Docker Desktop etc?

There are so many options, and Ingress isn't the answer for those. Ingress isn't a thing that can receive traffic, Ingress Controller is and it - itself has to be exposed somehow, on port 80 and 443.

The lowest common denominator is the one which works everywhere - kubectl port-forward. After that, hobbyists have a propensity for MetalLB, but it's not strictly safe at work with advertising ARP ranges.

I think the takeaway is YMMV. For production, it's of course Ingress Controller plus a cloud-managed LoadBalancer.

For anyone who is confused by the lack of context: the poster is showing us an issue with the GitHub UI which was circulated on Twitter today.

This issue makes it look like a commit in a fork is actually on the upstream repository - and you can also make a commit with any email address you like, for instance with Torvald's.

Disclosure: I'm the maintainer of the open source project - openfaas which this blog post talks about. It also covers other FaaS frameworks and why we see a "cold start" even when using Kubernetes as our platform.

> In this article, I’ll cover the reason why containers are slow to start - particularly on Kubernetes, then take you through our research and development that ultimately led to not needing cold-starts, or not needing to see them in OpenFaaS. There’ll be plenty of links and resources if you want to go deeper, and a FAQ towards the end.

Disclosure: I'm the maintainer of OpenFaaS, but also cover other frameworks and cloud functions (which I think are great for convenience btw).

The call for MIT / Apache 2.0 licensing at the end of the post was surprising, given the issues that the author outlined.

Grafana and Minio have both moved to AGPL / dual licensing from permissive licenses to stimulate licensing sales and revenue. As the maintainer of OpenFaaS, it's been a struggle to correct the balance of corporations freeloading [0] vs. contributing code or finances.

GitHub Sponsors has also given very mixed results [1]

In the same week we saw the maintainer of a popular OSS work almost throw in the towel [2]

After 5 years, we're starting to see customer interest in paying for OpenFaaS - but only by holding new features and enhancements back i.e. Open Core.

There are still problems with this approach, and paying customers end up funding all the R&D required for the free users - amongst other things like competition from hobbyist users who covet features that are more appropriate in production than on your Raspberry Pi cluster.

[0] https://github.com/openfaas/faas/blob/master/ADOPTERS.md [1] https://twitter.com/alexellisuk/status/1479870599853924352?s... [2] https://www.theregister.com/2022/01/13/opensource_apacheplc4...

It's now been ~ three years since I left my job and full-time employment to focus on OpenFaaS - on making functions easy to use and portable between clouds - https://blog.alexellis.io/openfaas-bright-2019/

Here's a bit about how it's going for me this year in 2022.

Hope you also enjoy the book recommendations if you're into personal development like I am, or are thinking of starting something.

It's been a busy year for the project and none of this could be done without the support of OpenFaaS' community and GitHub sponsors. Thanks to everyone.

Alex, Founder of OpenFaaS

Your job is probably triggering a fight/flight response constantly.

https://www.health.harvard.edu/staying-healthy/understanding...

Working with your hands and retreating to your shed allows you to sooth / calm rather than getting deeper into those stress hormones.

For me, it also gives time for reflection.

From the BBC: Men's Sheds: 'It's about so much more than making things' -> https://www.bbc.co.uk/news/av/uk-northern-ireland-49615371

Have you read Daniel Pink's When?

He talks about rhythms with most people being morning people - alert in the morning and ready for analytical work, entering a trough in the afternoon, then recovering in the evening. Are you mainly doing your shed work in the afternoon/evening, after work?