How to handle cloud resources in your application while running localhost

Hi (non-)binary people,

When you are developing an application and you run it on your own device, how do you do this when you have to rely on cloud resources like API Gateway, Cloud-map, S3 or a secret store like KMS? With RDS or with the DocumentDB you can run an addtional PostGresSQL/MySQL or MongoDB service on your device. (Although I mention several AWS services, I meant it in general.)

With for example the Application Load Balancer: I keep that configuration to a minimum and basically add configuration to send everything to a reverse proxy. Where the reverse proxy has the correct configuration, with retrying, setting headers etc. I understand that the ALB can do this as well (some maybe better or not, not relevant for this), but when I run my application locally, I still need to have something running to set these headers for example. As I can not set these, what am I doing locally then? I still want to know for sure that if I have a nice feature I can test this locally first before commtting and create a PR. Like I can not run a local ALB, but I can run a reverse proxy locally. So by keeping the configuration of the ALB to a minimum and use a reverse proxy, I can - if needed - run almost everything on my device.

How do you do this, I am eager to your experience.

"May the force be with you"

https://redd.it/qv8ynq
@r_devops

Prometheus alerts in grafana

Hi, all i have installed kube-prometheus-stack in a kubernetes cluster.

This helm chart provides a lot of alerting rules for kubernetes resources. This cluster is offline so i can't send alerts to an external destination using alertmanager. Instead the idea is to observe the cluster using grafana. Is there a way to see prometheus alerts in a grafana dashboard? I tried alertmanager-datasource but does not work maybe because is not updated for a while

https://redd.it/qvbdh5
@r_devops

How incidents made me a better engineer

My colleague has written a great post about handling incidents has made her a better engineer, and how people can make the most of that opportunity.

Think it's a great post for people in this sub-reddit!

https://incident.io/blog/incidents-made-me-a-better-engineer

https://redd.it/qv94gf
@r_devops

Buildpacks pack CLI for FaaS

Does anyone know if there is something similar to Buildpacks for FaaS?

Buildpack with pack CLI allows you to create Docker images based on the code only, which is great. Though, I'm looking for something that can look at your code and create Lambda (AWS) or a Cloud Function (GCP), or even better if you can choose the cloud provider where the FaaS will run to generate the code for you.

https://redd.it/qvjfhk
@r_devops

Need some guidance on building out monitoring/observability from the ground up

So I've semi-recently started at a new job as their only DevOps person where a lot of infra things are greenfield and I'm tasked with getting things built out. I generally am pretty comfortable with most areas of infra (including Kubernetes, been using it in prod since 2015), but I feel a huge gap of mine is monitoring/observability.

I've setup the many metrics/logging/alerting systems before in different environments (Prometheus, DataDog, Sensu, etc), but I've not really actually done much beyond that as usually other teammates in the past have taken on actually working with dev teams, identifying metrics to gather, and getting dashboards/alerting setup for those.

In this new job, one of the first big projects is to get monitoring going and to help the dev teams get started on this. I've already got Prometheus/Grafana setup, and they've got the client library in the main service going. However I'm a bit overwhelmed on how to help them beyond this point. They ask questions about Prometheus and monitoring that I'm not able to answer, and I don't feel equipped to lead here since I'm trying to fill in this gap of experience.

What are some good resources for myself as an infra person on observability/monitoring as well as Prometheus best practices? And are there good resources for me to send to them for them on how to effective use Prometheus to monitor their services?

https://redd.it/qvmrn0
@r_devops

What does the structure of your engineering or development team look like (i.e job titles/role responsibilities)? Ex. Does it include QA Testers or are they a part of a separate QA team? Which team would your SRE(s) fall under?

Trying to do conduct some DevRel research for the company I recently started working for, and would like to better understand the internal structure and dynamics of software teams. Any insights you could provide me would be great. Thanks!

https://redd.it/qvige2
@r_devops

How would you manage this problem?

Hi. I am part of a team and we are currently working on migrating old stuff and deploying new services to kubernetes. The problem is we are terribly under capacity, so many teams and different services moving their stuff to K8S and we are accumulating a lot of technical debt because time pressure.

We also have the problem that once we start a project it is never ending. There’s no sense of closure or (iterations). Every single day is a ping from someone asking to change or tweak something. (Multiply this by 10). So the work keeps accumulating, everyone wants to change stuff, try something new and we have no time to do things right or develop improvements.

Quite frankly I am already annoyed and burnt out. I’m my previous company our SRE team had a product owner, our backlog was properly prioritized and teams wanting to do stuff would go through that process first. That allowed us to allocate the right time for everything. But now the work is so unorganized, developers think they can dispose of us for anything anytime and we’re stuck with unrewarding work because everything is improvised.

How would you fix this problem?

https://redd.it/qvwca9
@r_devops

Are there any reasons to choose VMs over containers?

Most of the pros and cons give the weight as to containers as better than VM, and it makes sense. But are there any use cases where using a VM makes more sense?

https://redd.it/qvzrkm
@r_devops

First Junior SRE interview tomorrow, need help

I somehow lucked out and got set up for an interview for a JSRE position tomorrow.

What should I expect from the interview process for a junior site reliability engineer? I'm passionate about the position, but with only 8 months of system engineer experience and a few years of scripting (cmd, powershell) experience, I am worried that I will fall short on the technical side.

I've read the phoenix project, and I'm currently reading the Google SRE handbook as well. It's all really interesting to learn and think about, I just don't have much experience. Is there a best way to approach such an interview, or what would generally be expected for a Jr SRE to know?

https://redd.it/qw47bq
@r_devops

Is this normal for a junior position?

I started as a sysadmin and I'm currently working as a "Sysadmin that writes all the automation" role.

It was established that my experience knowledge lays around: linux, python, bash, ansible, puppet, docker, automation - and that I'm not very familiar with k8s/helm etc..

Earlier today, I interviewed for a Junior DevOps position, and just recieved an this assigment (And yes, I noticed it has tons of typos).

The company itself seems really good but I'm not sure what to do about this assigment. I was given 5-7 days and the email said it should take no longer than 4-5.

I understand everything I'm requested to do (apart from some poorly phrased sentences) but some of these things requires a lot of learning to do properoly.

Does this seem reasonable? Is this normal for Junior (or even experienced) DevOps to get such an assigment?

Thanks ahead!

https://redd.it/qw3vso
@r_devops

Personal project ideas for GitHub/resume?

I'm working as a DevOps engineering intern right now and loving it. However, I want to work on some skills/projects in my free time that I can use in the future as part of my resume and on my personal GitHub to show my work. Anyone know where I can get ideas for projects I can do?

https://redd.it/qw6232
@r_devops

Event exporter, observability something alternative for opsgenie stack

Hi
Do you know something better than this project https://github.com/opsgenie/kubernetes-event-exporter ? I use Opsgenie but I don't like it. Maybe there is better software that creates own webpage?

https://redd.it/qw9f1h
@r_devops

The current job is great, really, but I don't do any programming. How screwed am I for the next job search?

The next level up for me is the $200K+ senior bracket working remotely in a MCOL city. I work with Terraform, Kubernetes, and GitLab on a daily basis, but I don't do any programming, like with actual Python code, on the job. I know Python from previous gigs and have open source contributions but can feel my chops getting rusty. What would you recommend to stay up-to-date for when the next job search comes around? I'd like to avoid leetcode and the jobs that require it if at all possible.

https://redd.it/qw9axp
@r_devops

I'm having difficulty learning Go

Landed my first ever role as an SRE back in February. Previously I was a Networking and Systems Engineer in your traditional Enterprise, MS heavy (but not exclusive) environment. This new company took a chance on me and introduced me to the world of kubernetes, docker & CI/CD.

I already had a bit of python knowledge going in - but reinforced it by buying more books and doing the exercises in each chapter and all that jazz. Same goes for learning the fundamentals of kubernetes, docker, Git CI and Argo CD. I primarily use python or bash when i'm implementing automation - and although I'm no elite programming hero, the scripts get the job done and I try my best to keep my code clean. I did a rather heavy amount of ansible at my previous jobs - but most of it dealt with networking equipment and deploying out services to *nix machines - not nearly as expansive as the playbooks i'm writing now.

My boss asked me if I'm willing to learn Go. Of course I am. I'm willing to learn anything - but I'm having a tough time with it. It's so different it's hurting my head. I spent two days just trying to understand how the hell pointers work and even now I still don't think I've quite grasped the concept of it.

When I was learning python, the first book I read and went through was No Starch Press's Python Crash Course. That book was awesome, and very well written - and I picked python up relatively quickly and was able to write really handy scripts within a few weeks.

Not so with Go. I first picked up The Go Programming Language. First chapter almost made me want to cry. So i decided to dumb it down and get O'Reilly's Introducing Go. This book is significantly better, however it tends to gloss over some things that I feel need to be explained a little more for me. I kinda need hand holding with stuff like this. I never went to uni for CS. I went for Engineering and that was a long time ago. Are there some other books I haven't discovered yet that could take me through Go as if I was a complete idiot?

https://redd.it/qwe7uz
@r_devops

What is the easiest or best way to get my Django/React portfolio app online?

Major noob questions here. I....need help.

I've gotten the Django side of things covered and deployed, but it's my first frontend React app and I'm struggling a bit. I'm trying to host them both in a Digital Ocean droplet to avoid having to deal with CORS and CSRF complications. To be totally honest I was hoping I could just run `npm start` instead of building the thing because I'm new to react -- I willingly invite the vitriol this will invite. In fact, I think that 'npm start' is so (appropriately) inefficient that it's eating up all of the ram of my thrifty little $5/month digital ocean droplet and crashing. I desperately also tried heroku and some digital ocean 'Apps' service to no avail. My next step is probably just finding the absolute most responsible way to deploy a React App instead of this chickenshit approach, but again, the CORS and CSRF stuff is waiting to bite me in the ass if I deploy the React side separately from the Django backend.

I picked up docker today and I actually got Docker Compose working locally, but it freaks me out to add a MYSQL client to docker-compose, as configuring a MYSQL server is a bit of a reach for me to begin with. I am totally willing to invest in becoming stronger at Dev-Ops, but I just haven't focused on it much yet.

Essentially, I worked on this thing hard for two weeks and I haven't really left myself enough energy or focus for the deployment side of things. I could ration some time to getting it done right and really understanding Docker, and how it works with React and Django, but I'm not even sure -- might be overkill. I'm super worn out after gnawing on this all day. I'll wake up tomorrow with more initiative and intentionality. I am very open and interested to hearing any perspectives you experiences dev ops may have, thank you!

https://redd.it/qwf6im
@r_devops

How did deployment look like before Docker and Kubernetes?

Hi friends,

Just curious, what did deployment (and DevOps if by another term) look like back in the day before Docker and Kubernetes? Tried to Google for answers but all articles point to some third party Kubernetes/Docker services or bloggers so didn't really help.

https://redd.it/qwf0ll
@r_devops

Concept for an app/platform - code it myself or hire right away?

Hey everyone,

I have a very concise idea of an app I would like to create, very simple user input, small size, some social media elements. I know Python (background mostly in web automation) but I have not done anything in Kotlin yet which I would like to go with because of nativity on Android and the option of using a KMP to ideally make it less of an effort to deploy it on iOS.

Now I'm wondering: Should I build a first version on my own or hire someone right away to do it? I enjoy coding but I am sure it would take me longer than someone who knows his/her way around better; also because I do enjoy directing overall working processes and want to focus on that more (i run my own business with a few employees).

In case anyone has experience in this field: What would you recommend?

Thanks in advance!

https://redd.it/qwn83o
@r_devops

Install a Kubernetes cluster on Hetzner in 5 minutes (7X cheaper than AWS)

Having an out-of-the-box Kubernetes installation can be a big hassle, and I couldn't find a good step-by-step tutorial to set it up until I discovered this open-source project by u/Vitobotta.

I wrote a full article that goes over how to create a cluster using k3s on Hetzner – and how to set up monitoring.

The cluster has one master node and 3 worker nodes, for a total of 12 vCPU and 48 GB of memory – with a monthly cost of \~$54 (while the same config on AWS is $407).

You can read the full article here.

Any feedback is welcome :)

https://redd.it/qwpsna
@r_devops

Looking for a simple, managed cloud provider with CI/CD

Hi folks!

I've worked with the big clouds like AWS, Azure and Google and they're doing a great job, but I'm now **looking for something more guided** and limited.

The options of the big clouds are a bit overkill, depending on the user. As far as I know, Heroku and Digital Ocean are more what I'm currently looking for, but I'd like to get some more **recommendations from you guys**.

Some requirements:

* CI/CD (take source from GitHub and deploy automatically)
* Virtual machines (not managed)
* Managed apps (like AWS Elastic Beanstalk)
* Container (Docker, ...)
* Managed databases
* Highly scalable
* Enterprise support

Any recommendations?

**Thanks** in advance!

https://redd.it/qwqdl7
@r_devops

BLOG Unpacking Observability: The Observability Stack

I am sharing this article prepared by my colleague at Tucows that gives an overview of the Observability Tech Stack. The post is the second piece in their "Intro to Obersvavility" series.

Link here: https://storiesfromtheherd.com/unpacking-observability-the-observability-stack-93d4733e2a72

https://redd.it/qwsf0n
@r_devops

Why use the Docker image and Azure Registry, and not just deploy to Azure App Service?

what would be the advantages and disadvantages

https://redd.it/qwwwxf
@r_devops