What CDN fits this use case?

We have an e-commerce site with a long tail of xx,xxx product pages. Many of these pages are visited infrequently, but when they do get visited, we need them to be performant.

The application is built with Next.js and hosted on Vercel. The pages are statically generated at build time.

Vercel provides their own CDN, but it evicts items from the cache for being accessed infrequently, and their definition of infrequently is if the item “only” gets requested once per day, for example. It doesn’t seem like you can change the cache expiration time. So that doesn’t seem to be a good fit.

I’ve read that AWS Cloudfront provides more control over the cache expiration time, but unless I was reading outdated information, it doesn’t seem to perform well with manual purges. We deploy pretty frequently at times, so being able to purge and repopulate the cache quickly is a feature we’re looking for.

I’ve also read about Fastly, but my understanding is that it’s super expensive, and I wasn’t able to get a hold of anybody when I reached out to them with questions.

Is there a better approach to this, or another CDN I should be looking at? Or maybe I’m wrong in some of this analysis?

Thanks in advance for any guidance.

https://redd.it/1dpacb3
@r_devops

Alternative to Strongdm?

Strongdm just jacked our price by more than double in renewal. Wondering if anyone tested out other tools besides Teleport. We use it for accessing our EC2s, mostly SSH.

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

Github Actions view variable in a different workflow

Hello guys!
In my workflow I have an input variable which decides which kind of secret I'm using.
I would like to 'publish' this variable, so in the next workflow run, I will be using the same input variable, or overriding the input variables, instead of having to manually input it again, or without it defaulting to the default option.

This workflow is used by multiple people, and each one of them needs a different secret, so it needs to not default the new variable for everyone I guess.

Is there anything like this that exists?
The best solution i've got to, is 'publishing' it by passing it to the step name, and then taking it from there, but it feels a bit wrong.
for example:

name: test
on:
workflowdispatch:
inputs:
secretchoice:
description: 'Override secret id'
required: false
SECRETID:
type: choice
options:
- "secret1"
- "secret2"
- "secret3"
- "secret4"
default: "secret1"

push:
branches: ["master"]

jobs:
test:
runs-on: ubuntu-latest
steps:
- name: "set [ENVSECRETID: ${{ github.event.inputs.SECRETID }}]"

run: echo "published on stage name"


would love to hear an expert opinion

https://redd.it/1dpllp5
@r_devops

API connection with incorrect SSL only working on occasions


I have an API server that excepts post requests from the server of clients, which are sent via a wordpress plugin. However I noticed that for one request (where I used wpremoteget instead of curl) only worked for a hand full of users. Then I found that the source of the error was a mistake with my SSL certificate that I forgot to fix. It was towards my main domain and not the subdomain.

Does anyone know what the reason may be that this worked for a couple of users, but not all of them? I tried it from my local testing environment as well as from my staging site multiple times and it always worked. I only noticed the error, when I tried sending a curl post request via the linux console of my server.

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

Credits and resellers of $100k. Any risk?

As I was searching for how to get AWS/GCP credits, I bumped into multiple people who offered to resell credits. E.g. they’ve got $100k (or less) in credits, they’re not using them, want to split 70/30 or 60/40.

Thinking of taking up their offer after the credits I get run out.

1. Is there any risk if I join an organization?

One of those sellers is not comfortable with disclosing who they are, asked for crypto payments (red flags) but eventually said they’d accept PayPal. He also offered to make an advance of 10k which I can pay in 1.5mo from now.

Is there any risk e.g. can they see my user’s data? Or do I risk losing anything?

1. Can I just keep on joining different organizations, jumping from reseller to reseller?

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

User Behaviour Analysis / Mining

Hey everyone,

due to recent developments we are looking for a solution to monitoring user behaviour across an eCommerce-platform. We want to identify suspicious behaviour (potential data exfiltration, crawling from non-indexing sites, etc) and ideally get alerted when sessions seem suspicious.
We could do this based on access log analysis or session monitoring, but so far our research for tools in that regard has not turned up anything useful. We have monitoring for marketing (conversion, bounce rates) in place already and have things like dead links and infrastructure/performance problems covered, this is specific to user/session behaviour.

Wondering if somebody else is using tools for that purpose and has recommendations?

Thanks in advance :)

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

Any companies accepting positions (for visa holders 🙋‍♂️)

Sorry to ask, but want to see how markets are treating Visa holders in DevOps world right now. I tried two apply in LinkedIn and got instant rejection when I said I need future Visa Sponsorship (I figured its an automated system).

Cheers! 🫡

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

Has anyone used Velero for their backup and DR on kubernetes clusters ? If not what tools do you use ?

Velero is open source and is recommended by aws for kubernetes workloads but has anyone used anything different ?



https://redd.it/1dpshw2
@r_devops

First Interview - DevOps Student Position, Looking For Tips

Hello everyone!
I have my first interview coming up soon (as a 2nd year B.sc student of CS), and it's for a DevOps student position in one of the big companies.
I've spoken to the team manager as an initial phone interview, and she hinted at me that I'm not supposed to know so many things about DevOps (she knows I'm a student), and in the interview I'd be asked "logical" questions, and maybe some algorithms (in the technical part, of course that there is a regular, personal interview part).

In your experience, how should i prepare for the interview?
My current thinking is learning about the basics of DevOps (version control, Linux, CI/CD tools, cloud etc.)- but really only the basics of most of them, as well as practicing general interviewing, and grinding Leetcode questions. I'll mention I have less than a week until the interview.

Also, I'll use the opportunity and ask what do you think of DevOps as a student position? does it "lock me" in DevOps style jobs further on?
How can I take advantage of this first job to advance my skills as much as I can?

Thank you!

https://redd.it/1dpu9zq
@r_devops

Seeking Project Ideas for a Hands On DevOps Refresh

I’m currently the Head of DevOps at my company, which means I’ve been spending a lot of time recently leading people rather than building things. While I love my team and the work we’re doing, I’ve found myself missing the hands-on engineering aspect of DevOps.

For my sanity (and to sharpen my skills), I’ve decided to dedicate some of my free time to building and tinkering with projects. However, every time I sit down to start something, I end up staring at a blank screen, and my mind just goes blank.

I’m looking for suggestions on what to build or work on. Here are a few areas I’m interested in, but I’m open to any and all ideas:

1. CI/CD Pipelines: Always looking to improve and innovate here.
2. Automation: Tools and scripts that can streamline processes.
3. Cloud Infrastructure: I’m particularly interested in AWS, but I’m open to exploring other providers.
4. Monitoring and Logging: Enhancing observability with tools like Prometheus, Grafana, ELK stack, etc.
5. Security: Implementing security best practices in the pipeline and infrastructure.
6. Kubernetes: Anything from setting up clusters to advanced scheduling and scaling techniques.

I’m also very interested in contributing to open source projects. If you know of any active projects that could use some DevOps love, please point me in the right direction. I’d love to help out and collaborate with the community.

If you’ve worked on a project that you found particularly interesting or challenging, I’d love to hear about it. Or if you have any other ideas, please share!

Thanks in advance!

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

Cloud agnostic SFTP

What is everyone's favorite cloud agnostic SFTP solution for prod?

https://redd.it/1dpxp8z
@r_devops

Seeking Career Advice: Data Engineering vs. DevOps - Your Insights Needed!

Hi, Reddit community,

I need some guidance on choosing the right career path, and I’m hoping to get your insights. Here’s a bit about my background:

* I have a bachelor’s degree in Computer Science.
* Worked as a full-time Data Engineer for 1.5 years, with some minor DevOps tasks.
* Currently, I’m wrapping up my Master’s in CS while working part-time in DevOps, with some occasional Data Engineering tasks.

I’ve had exposure to a variety of technologies and found DevOps tasks particularly interesting. However, there are a few things that concern me:

1. **High Responsibility:** In my DevOps role, I often have to mimic the entire production setup in a testing environment, test changes, and then update the production. The fear of breaking something and the responsibility that comes with it is quite daunting.
2. **Constant Learning Curve:** There are tons of tools in DevOps, and I often find myself referring back to the documentation. It feels like I never quite remember how to use the tools as well as I did in my Data Engineering tasks.
3. **Small Teams and High Responsibility:** It seems like DevOps teams tend to be smaller, meaning each member has more responsibilities. Is this true in your experience?

With my Master’s nearing completion, I need to start looking for a full-time job. Given the current market, I’m trying to decide between pursuing a career in Data Engineering or sticking with DevOps.

**My Questions:**

1. **Should I seek a job in Data Engineering or DevOps?**
2. **Is the constant learning curve and high responsibility in DevOps something that gets better with time?**
3. **Are smaller teams the norm in DevOps, and does that mean significantly more responsibility per team member?**
4. **Any pointers or advice for someone looking to advance in DevOps?**

Your insights and experiences would be incredibly valuable to me. Thanks in advance for your help!

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

EU data protection (GDPR) question for contractors!

Hi!

I wanted to ask if it is realistic to work as a contractor for European countries, while residing outside of Europe.

Are there any workarounds to comply, and ensure full data protection for clients?

Will people be willing to hire me if that is the case?

Really appreciate anyone helping me out!

https://redd.it/1dq5id5
@r_devops

What is the way to find DevOps / SRE person for my SaaS project?

As the title indicates, I am seeking a DevOps /SRE to work on my SaaS project. What is the best way to find this person? Fiverr and Upwork haven't been helpful. Are there any other sites or places I can look?

More specially I need someone who can do the following:

* CI/CD implementation
* Data privacy
* Logging/Alerts attached to Staging and Prod environments
* Platform security
* Maintenance (incident management) and bug fixing

Current Stack:

- MERN Stack (MongoDB, Express, React, Node), Bitbucket (currently) - repository, and AWS (EC2,S3)

Feel free to DM me if you have any further ideas/questions, or if you fit this profile and want to chat more.

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

MLOPS Question: Deploying Model Ensembles BentoML to Sagemaker Alternatives

Hi, we are using BentoML and serving our model endpoint in an ECS accessible thru an API call. The model endpoint consists of multiple computer vision models (e.g. YOLO models) working together. Our use case is as follows

* should use only CPU VMs for now (GPUs are quite expensive, but it depends on the MLOps design)
* API calls can be sporadic (i.e. mostly single calls, random, etc.) but there are rare times within the day that the API will be called maybe 4-5x simultaneously)
* Predicting if an API request for the model can be somehow predicted as to when someone logs-in to the application, but not foolproof
* The endpoint can take 1-2 minutes to give a full prediction for an API call (since its CPU) but can be faster with a cheap GPU endpoint
* The user does not need the results to be instant. But it shouldnt be longer than 2-5 minutes.
* should not crash when you send concurrent requests
* cheap

Currently, BentoML on ECS (probably because of a very small machine) crashes when it receives multiple concurrent requests (maybe the RAM is getting flooded).

Is Sagemaker a good answer to this use case? Are there some alternatives to this? Thanks

https://redd.it/1dqd24q
@r_devops

Doubt on Terraform resources sharing between tf states

Just an imaginary scenario,if I define same AWS resource in three tf states (dev,prod,staging) as that resource is shared for using in all environments.If I destroy tf state or remove that resource in any one of the environments tf state ,Will that actually cause deleting that resource? How normally handle these type of scenario? Or we have to keep different statefile for shared resource?
If this question is dumb,pardon.am just a beginner🤝

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

For a DevOps role, how deep should I dive into learning PromQL?

Share your experiences! Is basic understanding enough, or is diving deeper a must? Let's discuss the importance of PromQL in the DevOps toolkit and how much knowledge is truly necessary.

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

I systemized a True-Fans-Based Growth Model to make launching new products easier for myself

I launch new products cyclically. Just an indie makers life. Some were very successful, and some less so. I've been in the industry for many years, and I've heard many stories of enterprises from my friends in the industry. I love reading about successful startup stories and adopting the best strategies for myself. I believe that learning from those who have succeeded by thinking outside the box and pioneering new marketing strategies is the foundation that allows us to break patterns ourselves

Some time ago I decided to start writing down ideas and strategies that I personally find most useful so that I could return to them and draw inspiration from them. As it often happens with personal notes, the idea evolved

I divided the product launch into 5 stages (before the product is made, when the MVP is being made, pre-launch, launch, and post-launch), identified the problems I most frequently encountered (and still often face despite a long history with product launching and management) and wrote down strategies to deal with them

What ended up in the final version?
Primarily solutions. Lists of universal possibilities that can be applied in various niches and markets and my favorite - case studies, which allow me to draw inspiration. Of course, I also included articles and books that I used for support

I tend to overthink every obstacle that comes my way and having a simple, clear strategy with examples of their application helps me a lot. Sometimes it's enough for me to read about how industry giants solved their problems to “unlock” the creative process which is key to a success.

If you want to check this out, here's the link: bigfootatemysandwich.com/l/hdvic

Maybe it will help some of you too. Would be awesome if you left a comment. I want to give the knowledge for free as much as possible and so the only reward we get is the impact we are able to give.

PS The file is also available in PDF format, but honestly, it is much easier to use in Notion. It's 48 pages. It's hard to read as a wall of text.

https://redd.it/1dqk6nz
@r_devops

Google maps api not diaplaying all data

Through an AI Html code (as I have no coding knowledge). The map is more or less working but I have some issues:
The places display full information. Actually picture, website etc is missing
Not consistent results. He has to lookup for a specific keyword but does not show up always the same places.
The maps behaves wierd. The places information stays up when you click away

Api enabled: maps embed api, maps java script, maps static api, geocoding api, paces api new

https://redd.it/1dqmea2
@r_devops

VM create by Packer, no IP assign and can not an IP

Hi friends,

I use this link for create Qemu ubuntu image by packer,

https://shantanoo-desai.github.io/posts/technology/packer-ubuntu-qemu/
after create an image I run it by this command



Virt-install --import --name Ubuntu --memory 8200 --vcpu 2 --vcpu host --disk Ubuntu,format=qcow2,bus=virtio --network default --os-variant=ubuntu22.04 --graphics spice --noautoconsole 
but when I run this command I see no IP assign to this VM

Virsh domifaddr Ubuntu
Name MAC address Protocol Address
-------------------------------------------------------------------------------
and I see this too



virsh dominfo ubuntu
Id: 10
Name: ubuntu
UUID: 93f6a9fb-e331-4a07-b7c1-c324771b3ca8
OS Type: hvm
State: running
CPU(s): 2
CPU time: 306.8s
Max memory: 8396800 KiB
Used memory: 8396800 KiB
Persistent: yes
Autostart: disable
Managed save: no
Security model: apparmor
Security DOI: 0
Security label: libvirt-93f6a9fb-e331-4a07-b7c1-c324771b3ca8 (enforcing)

I do know which part of this guide must change



https://redd.it/1dqrfu5
@r_devops

Creating a Custom AWS API Gateway REST Terraform Module

Hi everyone! I hope you're all doing well.

Has anyone worked on a fully custom AWS Terraform API Gateway module for a PRIVATE REST API? I'm trying to set one up, but the API Gateway (REST) has some of the most challenging Terraform modules and syntax I've encountered.

We're aiming to create a REST Private API with several endpoints that will forward incoming private traffic to other private resources.

For context, we're avoiding the messages/publisher/receiver architecture between microservices (as per the CTO's preference).

If any of you have experience with this, could you share some advice on managing communication between microservices? We currently use REST API requests between microservices, but we're open to other methods.

Thanks a ton!

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