Meli, a Netlify-like platform for deploying static sites

Hi there 👋

We used to host our sites on Netlify, but our eyes fill with glitter when we hear open source and self-hosted 🎉. So, we built Meli, which essentially is a Netlify alternative that lets you deploy static sites and frontend applications with ease, featuring per-branch deployments, web/slack/mattermost/email hooks, an API, and a way to manage organizations, teams and sites easily.

We built Meli on top of Caddy, a very powerful HTTP server. We've used Typescript, Node (backend), React (frontend) and MongoDB for the database.

It's a beta, but you can install super easily with Docker Compose: https://docs.meli.sh/get-started/installation

Check us out at https://github.com/getmeli/meli 🚀

Looking forward for your feedback 😀

https://redd.it/kb5l6m
@r_devops

Octopus deploy pricing target server definition

For those of you who use octopus deploy, what does deploy target mean in the context of licensing? I read their licensing 1 of 2 ways.

1. Deploy target means a target/server that is registered into octopus deploy via agent or via octopus deploy inventory mechanism

2. Deploy targets means you can't deploy to more than x servers at once but there is no limits total servers that can be deployed to.

I ask because while we have 100s of servers, they are only clustered in groups of 2-4 depending on the app. If the second licensing structure is what they follow, i can use the free edition. Otherwise I need to find an alternative.

https://redd.it/kb4snf
@r_devops

Red Hat UBI - updating or installing new packages

We have a docker file where we run a couple yum commands - yum update, yum install telnet.

The problem is that when we try to build the image from a Windows 10 machine via Docker Desktop commands, we get a message saying "This system is not registered with an entitlement server. You can use subscription-manager to register."

I tried this same build on a RHEL machine, and we were able to install telnet.

Does this mean it's impossible to build docker images on a Windows 10 machine where we install new packages?

Red Hat said it's not possible and directed me to this document: https://access.redhat.com/articles/2726611

That document is pretty dense, and I read through it but didn't specifically see anything disallowing what we are trying to do.

https://redd.it/kb32pa
@r_devops

How to find all resources in an AWS account?

Vantage published this guide on how they discover all resources within an AWS account. I regularly do consulting for folks and use this tool as a means for doing discovery over what my clients have in their accounts. I thought this group may be interested:

https://www.vantage.sh/blog/how-to-find-all-resources-in-an-aws-account

https://redd.it/kb8kgi
@r_devops

Weird out of memory errors

I've deployed k8s cluster into 2 node pools (one huge and one small but auto-scalable VMs) on DigitalOcean. The cluster runs an application (which has \~10 deployments) alongside with prometheus operator and AlgoCD. But last a few days our application becomes offline for a few minutes / a few times a day due to high resource usage.

The problem is that kubernetes deploys all jobs/pods on a single node, when node throws out of resource error and fails k8s reschedules all pods on another node then 2nd node throws out of resource issue then 1st node becomes online again and k8s schedules pods onto 1st node and then it throws out of resource error then 2nd node becomes online, ... some kind of weird loop of events happens.

I couldn't figured out why does this happens. I have not enough experience in k8s to fix this issue. So what could I do in this situation?

https://redd.it/kbaf3s
@r_devops

Student has a Question about Docker and K8s

Hi,

I want to learn more about DevOps and I was going to spend my winter break learning about Docker and K8. But apparently, I just found out K8 is dropping support for Docker.

​

Does anyone experienced have any recommendation on what I should learn and the path to it? I want to learn how to set up CI/CD and deploy a PERN stack website.

https://redd.it/kbb2vk
@r_devops

Is Docker still the way to go to for containers?

Hi DevOps friends,

A few days ago I chatted with a Director of a PaaS startup in hyper-growth mode and he told me that I shouldn't use Docker anymore for two reasons:

1. Kubernetes is deprecating Docker as a container runtime, see: https://github.com/kubernetes/kubernetes/blob/master/CHANGELOG/CHANGELOG-1.20.md#deprecation

2. He mentioned reproducibility issues with the archives (containers) and that Docker is aware of this common issue. Personal notes here: I worked for an ML/DL PaaS and never saw this issue in 3 years of works.

Instead, he suggested using OCI (https://opencontainers.org/), but actually, docker containers are OCI complaints! This is where started my confusion. Did he confuse docker runtime with containers?

From my understanding, even if you have a k8s cluster, you will not have any problem running docker containers given that they are OCI compliant.

I'm feeling a bit confused here. My guess is that we were talking about Docker (which is an entire tech stack!) from different perspectives, for me is still the best way to build and ship containers, but for him was only about the runtime.

Should we still use docker (at least to build containers)? Would love to hear what you think.

https://redd.it/kazzqi
@r_devops

What does it mean to design and build production level CI/CD pipeline from scratch ?

Apologies for such a simple question. But I always see this in job description. I am a Junior who has mix experience of bit of everything from Cloud to Kubernetes heck even some basic scripting but nothing too crazy that really set me out in the market. What are the tools and skills I need to build a production level CI/CD from scratch. ? I don't want to simply deploy "Hello world" app but something that will challenge me to up-skill myself.

https://redd.it/kbdglj
@r_devops

Frustrated of the company's decision

Hi, I am a devops engineer.

I am currently working in a company that is making a lot of profit during the coronavirus pandemic. We even had an increased in sales and retention this year.

However, recently the company decided to dismantle all the teams that are under the incubator group. The teams within this group are responsible for pilot projects or products with new POCs for about 5 years. An example of an output was a new feature that was created by a team that had automated a process that requires 20-30 employees.

I decided to leave the company because I do not see eye to eye with the company's decision. I can't do anything to voice out my opinion. We as a team automate a lot of business processes for some new products which could potentially be applied to the current products that we have.

From my perspective, if a company is making more money, should take the opportunity to take more risks to allow innovation. A company that missed such an opportunity, will lose out in the long run.

Please let me know your thoughts.

https://redd.it/kazs7d
@r_devops

The ethics of Pull Requests, being the “Reviewer”

Hi,

I wrote this blog post yesterday, the 2nd of a 3 part series. Maybe it could be interesting for this channel as well.

Link: https://werner-dijkerman.nl/2020/12/10/the-ethics-of-pull-requests-being-the-reviewer/

Please let me know your thoughts about it.

Kind regards,

Werner

https://redd.it/kb1vff
@r_devops

Is anybody can advice advanced Jenkins courses (shared library, pipelines, groovy)?

Is anybody can advice advanced Jenkins courses (shared library, pipelines, groovy)?

https://redd.it/kbgdv9
@r_devops

How do you implement API Gateway for a Microservice project?

I want to write an API Gateway for handling requests and routing them to proper services, it's kinda like a reverse proxy. But I want to add some functionalities like monitoring and analyzing requests before they reach final services. There are many people recommending tools like Kong, AWS API Gateway, Nginx while other people prefer writing one by themselves. So what are the pros and cons for using tools and writing your own?

Here are my concerns for this project:

* Scalability (easy to upgrade, add new features)
* Work well with Terraform and Heroku.
* Low cost, as much as possible (in terms of pricing and efforts)

Thank you!

https://redd.it/kbgczf
@r_devops

How do you do environment seggregation?

Currently we are running a separate Kubernetes cluster in a separate vpc for each of our envs, three dev envs, staging, sandbox and prod. There has been talk about collapsing into two vpc's and two Kubernetes cluster, just prod and non-prod. Then segregating environments based on kubernetes namespaces.

How do you divide environments?

https://redd.it/kaurvo
@r_devops

Change backend storage for hasicorp external vault

Hello,

I am using this terraform module to provision vault. Now, I want to change the backend storage. This seems to be not possible using the terraform that I am currently using ie https://github.com/terraform-google-modules/terraform-google-vault. If anyone can confirm this that will be helpful. According to the documentation we can use several backend storage storages. By default, it is currently using GCS as storage. I want to use MySQL.

https://redd.it/kaxup1
@r_devops

learn nodejs to write AWS Lambda functions

Hi,

I have a desire to learn nodejs so that I can write AWS Lambda functions. I use Python today but would like to learn node as I see plenty of AWS samples and examples in node. Looking for some pointers to simple tutorials where I can start.

Thanks

https://redd.it/kawjpf
@r_devops

how to uninstall argocd

how to uninstall argocd?

https://redd.it/kayuih
@r_devops

Do you think Kubernetes will keep on growing in DevOps land from 2021 on-wards?

Do you think Kubernetes will keep on growing in DevOps land from 2021 on-wards?

https://redd.it/kaqqxx
@r_devops

Do we have a consensus on where pipelines should stored?

Do you consider ci-cd pipeline as a part of the codebase?
Well, for public projects, especially open-source, I think it doesn't make much sense, yet Microsoft keeps Vscode build pipeline in the app repository. I am interested in the internal projects that you carry out in your company. Especially if you are using terraforming for immutable deployments. Do you keep stuff like terraform, packer, kubernetes manifest, dockerfiles and pipelines separately from your application repository?
Is there a consensus on best practice?

https://redd.it/kbnrd4
@r_devops

Thoughts on what belongs in a pipeline and what belongs in the build tool

Hello,

the question: how much should be done in locally runnable build tools vs build servers?

background:

I work on a project with an architecture similar to as follows:

- individual microservices are developed by semi-autonomous teams are part of a bigger project
- this bigger project has a provisioner which is used to install the project onto a cluster of VMs. It installs a variety of technologies such as docker, kubernetes, etc, to these VMs.
- this provisioner is released as a tarball with all necessary build artifacts, which all have fixed versions themselves. this is done because some of the install targets are not available through the public internet
- configurations which are included in the provisioner are versioned with the provisioner

this should give a rough idea of the approach, which I am sure is very common; individual sub-components product build artifacts which are orchestrated via a provisioner

the CI/CD challenge becomes how to automate things such that developers and operations engineers don't need to manually collect and version the provisioner each release. originally, this was how it was done, and it was a very slow, error prone and stressful process which resulted in lots of over time and sleepless nights.

I threw together some jenkins pipelines to automate most of it and it's been working well enough, but I'm not all that happy with it myself.

the issue is a lot of the build process exists in pipelines which are run in a jenkins server. these pipelines are under version control as well but they do an awful lot. building our compiled artifacts, dockerizing them, performing sonarqube and twistlock analysis and updating our provisioner all happens within these pipelines.

we use maven for some of our projects, and what I like about maven is it can be run locally by developers. there are plugins which can build docker images, perform sonarscanning and such. It's current quite a specialized task to update our pipelines, and I can't help but wonder had we done more in maven and less in jenkins if we might have been better off.

the issue with maven is it's got first-class support for Java, but kind of leaves our other projects out to dry. Is there a more language agnostic build tool which could be run locally that could be used to define our build pipeline? is this something other people have found is better than the Jenkins, GitHub Actions, TravisCI heavy approach that we're using now?

Interested to see what others think on this.

https://redd.it/kbr5wv
@r_devops

DevOps/SRE questions

I just wanted to share my experience. I went through couple of interviews for DevOps Engineer positions.

**Offer 1**
* Base: 89K
* Bonus: 7%
* Location: East Coast

**Offer 2**
* $61/hr on a 6 month contract
* Location: East Coast

**Offer 3**
* Base: 75K
* Bonus: 10%
* Location: South

**My Skills**
* AWS
* Python
* CICD

I did not take any offers because either they were less than what I make currently or not worth a swich because the increase is minimal. Have salaries taken a hit due to Covid? I have 2 YOE. I am hoping fellow professionals who have interviewed for similar positions can share their experience.

Stay safe everyone

https://redd.it/kbuigv
@r_devops

What factors do you use to determine your 2021 DevOps forecast ?

Example - Delivery Lead Time forecasted for the business in 2021

https://redd.it/kbuit8
@r_devops