Understanding self-hosting a public load balancer for small/early project infrastructure

I'm interested in self-hosting a public load balancer for my infrastructure supporting small/early projects.

I know, I know: It's stupid, I shouldn't do it, I'm insane, stop and use AWS services.

Assume I want to learn the hard way and that's not an option here.

With that said.... I'm curious, generally speaking, what kind of capabilities I'd have, say, self-hosting HAProxy on a t3.small instance (or maybe even a t3.micro)?

AFAICT, HAProxy is wildly performant and can do alot with little. And I'm not doing alot.

As far as average capacity, I'm planning for 'a few hundred' users of an unexciting, JSON-serving mobile application backend. HTTPS, but no WebSockets (yet), file transfers will be routed through S3, nothing complex otherwise.

As far as spike planning, I'll be using pre-configured AMIs for my infrastructure and I'm OK with an autoscaling SLA of 'a few seconds' to horizontally scale up the load balancer, should that ever happen. But I'm not really planning to have spikes and, if they happen, I'm tolerant of some short downtime for horizontal scaling.

As far as ongoing capacity planning, I'd like to be able to have some stats I can keep an eye on and, if the host resources start to saturate, manually horizontally scale as needed (I think CPU/memory/networking probably works for this).

Again, this is for my personal cloud, personal projects. I'm in for the learning.

Is this all a reasonable perspective? Anything I'm missing?

https://redd.it/1dahpk4
@r_devops

Employer seeks my input on the job posting for the position they plan to promote me to.

First off, I'm pretty excited for the potential promotion. I'm a hourly tech support employee at a company that I've worked at for roughly two years. They want me to fill in a configuration management engineer role. I've somehow convinced them that I'm the guy for the job, with zero experience in configuration management.

My technical director sat me down, and requested my input on the job posting for this role. I though this was odd, but am I wrong to assume that they're requesting my input because the know absolutely nothing about the role? All they, and I, know is that they need it. I'm thinking that this could give me some leverage and I need some opinions on how to go about this. I want to be able negotiate a salary (which hasn't been decided upon and is negotiable) and set expectations.

The resume requirements were bare minimum, something I'd expect from the current roll I fill. These include:

1. IT related Bachelors and/or relevant exp.
2. Self-starter
3. Problem-solving skills, communication, documentation skills

The job description excludes skills, knowledge, and expertise in things such as:

* Basics of containerization and the orchestration of containerization
* Basics of YAML/Declartitive configurations 
* Basics of networking including – OSI layers, protocols, IPs, DNS, gateways, routes
* Operating system knowledge (Windows, Linux)
* Terminal proficiency (in Bash or PowerShell )

Would you add all/some of these to benefit the salary? Leave some out to temper their expectations? Let me know your thoughts.

https://redd.it/1dahg76
@r_devops

Best masters degree for DevOps?

Not sure how common it is to pursue a higher level degrees, but if I were, what would everyone suggest?

From what I can tell there has been more emphasis on being able to program at the enterprise application level. So that immediately makes me think a computer science masters.

Would love to hear everyone’s thoughts!

https://redd.it/1daktkb
@r_devops

GitHub Actions Not Capturing Secret

Hi, I've encountered a very frustrating issue when trying to automate deployment using GitHub Actions. It should be pretty simple...build the Blazor site in a release config, install Wrangler, and be on our way, but it's been causing me a great deal of trouble for awhile. Here is the current code I'm using:

name: Deploy to Cloudflare Pages

on:
push:
branches:
- main

jobs:
deploy:
runs-on: ubuntu-latest

steps:
- uses: actions/checkout@v2

- name: Setup .NET
uses: actions/setup-dotnet@v1
with:
dotnet-version: '7.0.x'

- name: Publish Blazor app
run: dotnet publish -c Release

- name: Install Wrangler
run: npm install -g wrangler

- name: Deploy to Cloudflare Pages
env:
CLOUDFLAREAPITOKEN: ${{ secrets.CLOUDFLAREAPITOKEN }}
CLOUDFLAREACCOUNTID: ${{ secrets.CLOUDFLAREACCOUNTID }}
run: |
wrangler pages deploy ./bin/Release/net7.0/publish/wwwroot --project-name myproject

However, it fails, saying it's expecting a CLOUDFLARE\API_TOKEN environment variable. I've tried all kinds of things, but I can't actually seem to access that secret. And of course, I can't log it. I've tried asking a couple of AI assistants, but their guess seems to be as good as mine. If there's something obvious I'm missing, I'd love to hear.

https://redd.it/1dalf36
@r_devops

Mechanical Sympathy approach- does it still hold up?

"Mechanical sympathy", is a term attributed to Martin Thompson that means the better a developer understands the hardware, the infrastructure, and how things run, the better they will be at coding and avoiding and identifying problems down the line. Originally, the term comes from race car driving and it reflects the driver having an innate feel for the car, so they are able to 'feel' how to get the best out of it.

We referenced mechanical sympathy here: https://www.getambassador.io/blog/empowering-developers-with-guardrails, but we'd be curious to hear others thoughts on it.


What do you guys think? Does it hold up in how you lead your dev teams or is it comparing apples to oranges?

https://redd.it/1damjk6
@r_devops

Dev Wants To Use Unity

The software is supposed to enumerate Microsoft SharePoint permissions on a tree diagram on the web. Originally we were using D3 but had some spacing issues with the nodes and overlapping text.
So the developer gave up and is playing around with Unity saying that he can compile it into assembly and overall it allows for more control. I.e. custom right click menu. I am not a developer but what are the pros and cons of doing it this way? It just seems uncommon and I get the feeling we are going to run into issues in the future when we want to implement more features. I know he tries to go back to unity any chance he can get, even for mobile app dev at one point. Thanks for anyone's insight!

https://redd.it/1dancba
@r_devops

Anyone heading to re:Inforce and interested in access/permissions control?

Hey, we're gonna be at re:inforce in philly. Just putting out feelers to see if any one is dealing with least privilege policies / SCPs / interested generally in cloud permissions control. We wanna hear from real people what their pain points are.

https://redd.it/1daje0p
@r_devops

What constitutes a 3yoe vs 5yoe 7+ yoe in DevOps?

I have been boggling how experienced vs junior positions looks with respect to technologies worked on and extensiveness of the work?

https://redd.it/1dauy1x
@r_devops

Azure devops question

Legit question...is Azure a cult? I worked at a global company and was in a server management role and things were 'normal' seeming, but was then asked to join the azure team in the company. And in the beginning of it all, it had the vibe of 'welcome to the team, glad you're here, everthing's fine!' I did everything that was asked of me while still taking some liberties on what i thought was best to implement for the way forward. As things progressed with my team members things got a little.....strange. I was beginning to not feel safe and was amongst people that was into some pretty gnarly shit.
I eventually had to leave the company i worked for for 10 years. I decided to move to a different company that seems to be the same shit with different sprinkles. I took an IaaS engineer role still in azure and I am seeing the same warning signs all over again. Is this the sign of the times? Is it a cult? Is it corruption? what is up with the type of people that this platform attracts where people think that they are 'god' or a 'plague' or talk about nuclear explosions or the power of suggestion or magic? It's really weird and creeps me tf out. Can someone please clarify so that i can better understand. I would like to believe that things are not THAT serious...but I've worked on 2 different clouds now and it seems to be the way of things... If so, I think it's time for a career change...

https://redd.it/1daxqn7
@r_devops

Learning systems performance monitoring, alerting and tuning(web server, database)

https://www.reddit.com/r/devops/comments/11h343b/how\_to\_learn\_system\_performance\_as\_a\_beginner/

I was reading this question. And the suggestion was to selfhost. So, I purchased a domain name and started to self host my website.

I've now deployed my website to production and installed zabbix in it as well as netdata. Now, can anyone tell me what path should I take?

https://redd.it/1dayu92
@r_devops

Looking for a git visualiser/simulator

My team is currently looking at improving our git strategy. I'm our only DevOps Engineer so it's mostly fallen to me but I'm also pretty inexperienced. So want to use this oppurtunity to get familiar with different strategies etc. I've been reading up on popular ones but also just want to mess around and gain a better understanding of things

Is there a git simulator out there that will let me test things at scale? Maybe something that will let me choose or implement a strategy, add "devs" and then simulates a couple months or something?

https://redd.it/1db1vwr
@r_devops

Join the Microsoft Azure Fundamentals Challenge

🚀 Microsoft Azure Fundamentals Challenge 🔹

The challenge has begun! Join now and unlock the power of Microsoft Azure AI resources for your future projects. The learning adventure continues until June 16th! 🗓🌐

It will be an incentive-driven competition, where participants who rank highest on the leaderboard will qualify.



🎁 Perks Await You:

🌟 LinkedIn Premium Voucher (12 months)

📜 Certificates of Achievement

🎖 Exclusive Badge



📝 Registration Form:

Please take a moment to fill out this form and incentives will be sent to the email you provide. The email should be you given one for the event registration.

https://forms.office.com/r/Ki1uwJS4yM

🔗 Challenge Link:

https://learn.microsoft.com/en-us/training/challenges?id=8d3f19e5-612e-4cb9-ad8e-5eccd2ce68ba&WT.mc\_id=cloudskillschallenge\_8d3f19e5-612e-4cb9-ad8e-5eccd2ce68ba&wt.mc\_id=studentamb\_209465

🚀 Don't miss out on this incredible opportunity to learn and earn fantastic rewards! 🙌🏻

#Sweepstakes #Azure #Microsoft

https://redd.it/1db6fdn
@r_devops

Industries that hire DevOps Engineers

It doesn’t seem like all industries have DevOps positions. They seem more prominent where software tends to be the primary product, which I can understand. What other industries hire DevOps? And if a company doesn’t use that terminology, what are some other terminologies they use, besides Cloud Engineer (which I understand is a different role)?


https://redd.it/1db6wck
@r_devops

Reducing "All Hands on Deck" Incident Culture

I work at a company that tends to have an "all hands on deck" incident culture, very similar to this post from r/sysadmin. My team owns an application networking infrastructure platform that has internal developer teams as our customers.

We have built our platform to be almost completely self-service to developer teams, including providing telemetry tools (logs, metrics, traces) that developers can use to troubleshoot problems that pass through our platform.

However, our team has noticed a pattern during incidents where a developer team and/or their leadership will pull our team into incident troubleshooting meetings if our platform is involved at all in the impacted business/application flow (ex: Client -> Platform 1 -> App 1 -> Our platform -> App 2).

As a result of this, our team has become more efficient on ruling out our platform from the problem (using the same self service telemetry tools mentioned before). For 95+% of incidents, our platform is not the problem and it ends up being an issue with another developer teams application. This efficiency has saved us a lot of time and avoided blame for incidents, but the constant pull for our team still adds up and results in a lot of wasted time on these incident meetings.

From a business perspective, leadership wants us involved whenever possible to reduce business impact and MTTD/MTTR. We are struggling to find a way to change the organizational culture where developers will use our self service tools to troubleshoot incidents rather than relying on us to interpret the telemetry for them.

Have you all had similar experiences at your companies as platform owners? If so, how did you improve this culture so that your team reduces toil and gets more time to work on platform improvements/features?

https://redd.it/1db8ynx
@r_devops

Had an interview for this "Optical Network Engineer" role. Was told I will automate a lot of physical and network infrastructure. Is this a good position if I want to get into DevOps?

This is from a multinational company that is big in the mobile and telecommunications industry. I was asked a lot of questions regarding linux, algorithms, python, databases and AWS. However, they spent 30mins more than my scheduled interview time and had another candidate waiting so they had to end the interview. Howver, they said we'd do another one later. I didn't have time to ask about what my day-to-day would look like. So now I wanted to ask if a role like this would be good for a DevOps career? Or if I should look elsewhere



Responsibilites
> Perform OS/NMS installations, upgrades, migrations, and network transformations.
> Perform Lab testing, Field Trials, DryRun, Proof of Concepts, Preliminary Acceptance Tests, Customer Demos/Workshops.
> Define test cases based on customer specific scenarios and develop MOPs and test procedures.
> Extend and expand automation delivery footprint, also to develop scripts and tools to automate the manual and routine tasks.
> Perform extensive lab work, both in person and remotely.


Requirements:
> Bachelor's degree or higher in Computer Science and/or equivalent work experience in Telecommunications engineering discipline.
> Demonstrable skill in scripting and programming languages: e.g., Python, Java, JavaScript, and good knowledge in algorithms & Agile development processes and DevOps practices.
> Strong Linux/Unix shell networking knowledge, as well as, Cloud, Docker, OpenStack, Kubernetes, Databases, Containerized Applications, Microservices, etc.
> Willingness and desire to do hands-on lab work, and are careful and thorough, with attention to detail.
> Ability to work via own initiative, you have to be self-motivated with flexible, positive, and creative attitude. Also, you should have a strong communication, documentation, collaboration, and interpersonal & Team Skills.
>
>

https://redd.it/1dcmzlj
@r_devops

How are you managing your s3 buckets?

Scenario, we have a couple hundred S3 buckets to manage and find using cloudformation a bit of a pain.
I was wondering if anyone has any alternative suggestions, looking into terraform and open tofu at the moment.

https://redd.it/1dcj4y6
@r_devops

ChatGPT/Groq/LLM in your terminal?

I'm working on an open source project called SheLLM, which is still pretty experimental. It adds ChatGPT or Groq (self-hosted LLMs coming soon) to your terminal, enabling them to execute commands (after approval), explain your terminal's context, and more.

The reason I started this project is that I often forget many parameters (especially positional ones) for various tools, so I find it really helpful. It can also write snippets, which I use daily, as my aliases and functions have become quite extensive.

Being in the DevOps world, my questions to you are:

1. As DevOps professionals, what are your biggest pains when using a standard terminal?
2. Do you think having "template"-like workflows in your terminal would make your life easier (not scripts with strictly defined and hard-coded variables, but more flexible automation)?
3. Do you often expose secrets/PII/other sensitive data in your terminal, or do you edit it in a safer way?
4. How do you think an LLM can help you in the terminal?
5. How frequently do you use aliases and functions in your terminal, and do you find them becoming unmanageable?
6. Are there specific commands or tools where you frequently forget the syntax or parameters?
7. How do you currently manage repetitive tasks or commands in your workflow?

Your feedback would be incredibly valuable for the development of SheLLM (maybe drop a star?). Thank you!

https://redd.it/1dcwk1r
@r_devops

Platform engineering on Kubernetes

I’ve been designing and implementing cloud platforms for a few years and put together a blog describing tenants of good design https://piotrzan.medium.com/how-to-build-cloud-native-platforms-with-kubernetes-1f0901a63a04

https://redd.it/1dcukok
@r_devops

from Software engineer to Devops Engineer

Hello folks,

I'm a senior software engineer (8 years experience if we can consider this senior :p ) I work as a fullstack Software engineer mainly on web application ( java ,spring and React/Angular stacks) in my last project with my client I have to manipulate kubernetes to deploy the app in all the environments and I started thinking that I can add the devOps skills to my resume. The problem is now I am not sure what Certifications I need, I'm thinking doing CKA and moving to AWS certifications in order to get a DEVOPS job.
I'd love to hear your thoughts and recommandations about this.

Thanks a lot for your time.

https://redd.it/1dcfz03
@r_devops

Struggling to get a single interview (USA)

I got laid off few months ago and since then been actively applying, soon I'll be out of all my savings. Really worried about things. There is a complete lack of response, normally your disappointment in job search would include some ghosting after interviews and some rejections but it’s radio silence this time around. I'm now even applying to local jobs that are asking for office/hybrid. The only thing I can think of is that maybe some hiring managers are turned off by me having no Linkedin? I have 3 years of experience and two proper engineering jobs under my belt.

I have also started to just look up companies and find their postings on their career websites in addition to indeed / LinkedIn but that didn’t help either.

Can someone please shed some light on the current state of things, is anyone getting hired? I think I'd be glad to hear someone got a job offer at this point because it would mean things aren't as bleak. Located in U.S.




https://redd.it/1dctdbe
@r_devops

Optimizing log management with AWS OpenSearch at Chase UK

Hey everyone!

Eugene Tolbakov from Chase UK gave an interesting talk about effectively implementing and managing AWS OpenSearch clusters at the last London Observability Engineering Meetup.

I thought some of you might find this interesting.

You can find the recording here: https://youtu.be/NWrrqRmDa20?si=-haj0rxpjI838JrP

Btw, if you're interested in Observability, make sure you join our Slack Community!

https://redd.it/1dctd8v
@r_devops