Reddit DevOps
270 subscribers
5 photos
31K links
Reddit DevOps. #devops
Thanks @reddit2telegram and @r_channels
Download Telegram
CertBot Lambda

Hi mates, i need using certbot inside Lambda function. Actually i found some workarounds but it is not working clearly. I need use certbot's all functionality. Btw it may not be in only just a fuction.

https://redd.it/q07hyn
@r_devops
Has anyone here automated the whole billing alert/notification/monthly report in Azure?

I am looking to do this, but want to source some ideas on how you approached it

https://redd.it/q15jjr
@r_devops
GitHub - debarshibasak/awesome-paas: A curated list of PaaS, developer platforms tools to emulate PaaS on cloud, Cloud IDEs and ADNs.

The PaaS, developer platforms have blown up in recent years. Lately, I have been curating and tracking these platform. Thought I would share with the community. I am actively maintaining it.

https://github.com/debarshibasak/awesome-paas

https://redd.it/q138sh
@r_devops
Is there a way to use the Jenkins core 'archiveArtifact' with Artifactory?

I really like the archiveArtifact call since it shows up pretty well on Blue Ocean. I was wondering if I could somehow use that with uploaded artifacts on Artifactory? It doesn't seem like it, but does anyone know if there is another way?

https://redd.it/q18htf
@r_devops
Dockers and Kubernetes for Data Scientist-MLOps



Please check out my article

MLOps-Dockers and Kubernetes Essentials for a Data Scientist

MLOps-Dockers and Kubernetes Essentials for a Data Scientist

and my previous articles

MLOps End-To-End Machine Learning Pipeline-CICD

MLOps End-To-End Machine Learning Pipeline-CICD

10 Amazing MLOps Learning Resources

10 Amazing MLOps Learning Resources

https://redd.it/q19j39
@r_devops
Chat-application using Jabber based on phone-numbers without centralized authority

Hello,

I have a Thai girlfriend. I want to use Jabber with her. She's loving and caring, but she's not always super intelligent as she struggles to understand things like the TCP 3-Way Handshake.

Her country is in a dictatorship and we both have 2 Jabber-accounts that work fine for chatting. I would like to have a system that allows us to register our phone numbers publicly with our Jabber accounts.

If for some reason, the government is able to block our Jabber account, we can just register our phone number for a new Jabber account on a new server and continue chatting. The only problem here is... how could we ever make this possible, without any centralized authority?

Would a system to register your own phone number ever be possible without a centralized authority that could be banned by the local government?

https://redd.it/q18ea7
@r_devops
IoT device connection using MQTT

Recently, I am facing a problem that IoT-enabled devices are taking time to connect to the MQTT broker. When the number of devices was little, it was fine. But when the number of devices increased, it took time to connect to the server-side. What could be the reason? I am using a mosquito broker. Is this a broker issue or anything else? If it's a broker issue, for industrial IoT which broker will be good for MQTT?

https://redd.it/q1242d
@r_devops
Running python code remotely and simultaneously without needing multiple computers

I have python code that preprocesses CSV files inside zipfiles and sends the data into an influxdb database on a digitalocean linux droplet using line protocol. I use Spyder locally to run this code on my computer to send info into the remote database. I need to give a category tag for each zipfile so I need to run the code on each zipfile one at a time unless the zipfiles are in the same category.

How would I set something up so that I could upload the zipfiles somewhere remotely so that I don't need to use my local computer to do the preprocessing and uploading process, and so that I could run the process on different files simultaneously? Is there something on digitalocean or AWS that would be best for this?

https://redd.it/q0mdbz
@r_devops
Shift left responsibly for CI/CD to developers?

Who did this already and why? If you did that, what worked best for you and what didn’t?
If you haven’t done that what was the reason?

https://redd.it/pws8a2
@r_devops
Resume for a remote DevOps Position

Hello everyone,

I am looking for remote devops positions but I haven't been very successful. I want to upload my resume and see if anyone can point out ways to improve it.

Also I got a local offer for more than a 100% raise. Is job hopping after only six months considered bad for your future career.

Thank you.

https://redd.it/q1fzj7
@r_devops
fresh out of college graduate recently hired at a company. i've always used source control, CI/CD and have been interested in SCM. I'm trying to encourage my workplace to use these tools and I never thought I'd have to deal with culture change from a dev perspective.

I've recently started working as a junior developer at a company - dream job and all.

However, I've noticed that there's absolutely no structure within the dev department. Everyone codes like spaghetti - nobody follows coding conventions and everyone just pushes all their change into one branch without a code review process.

Their reasoning was that most projects, devs work alone so there's no need to implement these standards or tools. Note the emphasis on "most" and that they still do collaborative work, but it is rare and I reckon it has to do with the fact that each dev is doing their own thing.

I tried to give suggestions and because I'm a junior, my suggestions often get shut down by devs with more seniority. At the same time, I don't want to impose myself on them (hiring for devs is quite hard as it is).

It seems that the philosophy with this company's devs is that they're here to deliver products that look nice on the front end fast, but it's pretty clear that because there's no standards that it's taking more time than it should to develop because every dev ends up working solo and no one can help each other.

Has anyone had experience with culture change in dev department? Any advice to give to a junior? Could I be approaching this the wrong way and that all this push from having more standardized source control processes to even applying design patterns isn't necessary?

https://redd.it/q1nyqn
@r_devops
Certified Kubernetes Application Developer (CKAD) Study Guide ebook

Developers with the ability to operate, troubleshoot, and monitor applications in Kubernetes are in high demand today. To meet this need, the Cloud Native Computing Foundation created a certification exam to establish a developer's credibility and value in the job market to work in a Kubernetes environment

The Certified Kubernetes Application Developer (CKAD) ebook is different from the typical multiple-choice format of other certifications. Instead, the CKAD is a performance-based exam that requires deep knowledge of the tasks under immense time pressure.

Certified Kubernetes Application Developer (CKAD) Study Guide ebook

https://redd.it/q1nuyd
@r_devops
Which tool(s) should I use to deploy containers?

Nobody in my team has any experience with Containers (Docker, Kubernetes etc.). I'm responsible for introducing them as we will like to use containers and everything that comes with it, such as Orchestration. Only Docker is set so far.

Our scenario is basically to set up multiple containers (at the moment 5 per client), which need to communicate internally as well as externally through APIs. My first POC includes Docker only, where I make use of Docker Compose. For every client I can simply run the docker-compose file and everything is set up automatically (as of now, we have a single server, that runs every container) in a few seconds.

The number of containers will grow during the next years and therefore I was questioning myself, whether we need an orchestration tool, such as k8s, or not?

Some more questions:

1. How do we monitor these containers (e.g. uptime/availability or container specific logs such as database monitoring)? Do orchestration tools provide these features or would we need external software?
2. What about fault tolerance/uptime)? Are orchestration tools better in this context (all I know is that I can advise Docker in the docker-compose file to restart a container when it isn't running)
3. As we don't need to scale (and likely won't need in the future), might an orchestration tool add to much overhead?

Are there more lightweight orchestration tools? Kubernetes seems to add to much overhead and complexity. We don't have large clusters running and we don't need to scale. Can you recommend Docker Swarm? Alternatives? Would a combination of Docker Compose and Portainer be sufficient?

As I'm fairly new to this as well, feel free to point out what I will probably need from an orchestration tool, but I didn't mention above.

https://redd.it/q1uxjf
@r_devops
AWS Cost Optimization by sharing resources: Is this relevant to you, and for which use cases?

We're currently exploring ways in which we could lower our cost on AWS per application environment.

We run a relatively high number of apps in a 3-tier ALB-ECS-RDS architecture. When it comes to staging environments & PR preview deployments, we must ensure these are deployed in a way that doesn't burn 50$ per environment a month.

The most well-known cost saving advice is reserving capacity on AWS (i.e. Reserved Instances). I've written a post about another method we're implementing, which is making sure we use every bit of the resources we provision on AWS. We have done similar things for RDS databases (where low workload deployments share an instance) and for NAT instances (which you need to make reliable yourself, but then they're way more cost-efficient than a NAT gateway).

I'm interested if this is something that can be considered standard/good practice, and if you have other use cases for sharing in mind.

https://redd.it/q1uqv7
@r_devops
What used to seem difficult and is now trivial

Hello,

More of a just an informal survey, but what was something that seemed daunting when you first started working DevOps and now seems trivial?

https://redd.it/q1vl6h
@r_devops
Hiring: Are we doing it wrong?

So our micro organization (\~20 employees total, tech team of 4) is looking to hire our first DevOps role. Up until now we've had a programmer (me) filling the space but you can imagine slogging through it is taking away valuable programming time. Not to mention not having the experience to get it right as quickly.

Anyhow we're finding it really difficult to bring folks on. I understand it's a very hot market and I think many of our candidates are simply getting snatched up by larger organizations. We can only control so much and I want to make sure I'm doing everything I can to make the process fair, useful, and ultimately find a good fit for both parties.

Our product manager is starting with a quick phone call to make sure they're understanding the general role, a bit about us, and feel out if they're a good personality match. Then we move on to a technical interview in which my canned request is:

>Hey there $FirstName we would like to schedule a technical interview this week. We will chat about your background, DevOps in general, and I can tell you the technical details of our setup. We will end with me answering any questions you might have. A small technical competence project will follow our discussion that you are free to do at your leisure. I wouldn’t ask you to spend any more than an hour on that portion.
>
>How does $DateTime work for you? Happy to reschedule off hours or on the weekend if you’ve got things going on during the day.

Our 1h call is a fairly surface discussion about their experience, what we're dealing with here at a technical level (showing our TeamCity, Octopus Deploy, and CDK steps), and just talking shop. I am showing our architecture* diagram and mentioning I'd like them to implement it and I'll send details after the call.

My email after the call is this, with the diagram included*:

>Hey there $FirstName thanks for joining the call. As discussed there’s a short exercise we’d like you to complete. This is a real problem we have and would like to evaluate your approach on it. To be clear we have already solved this – we are not looking to benefit from your work. Also we are not asking you to spend a whole lot of time on it – maybe just see what can be done within an hour. If your solution is a working one that is a bonus, but having it in a runnable state is not a requirement.
>
>Given this diagram below, in a platform of your choosing (TerraForm, CDK, CloudFormation, bash scripts, etc), we need to stand up and tear down these components within AWS in a repeatable way. Obviously you haven’t got access to the application files or any domain knowledge so we are only looking for the very basics of these components (a CloudFormation construct, a database, EC2 instances, etc). If it helps to pick a subset of the diagram then do that. 
>
>You can simply respond to this mail with your solution, and feel free to ask for any clarification. There is no time limit – feel free to do it at your leisure.

Out of seven candidates that have reached this stage only one has responded. They withdrew their candidacy the next day. Are we the problem? Would love to hear your honest feedback.

Thanks!

* - unfortunately I can't share the architecture diagram but it consists of CloudFront, ALB, AutoScaling, ElasticSearch, and RDS.

https://redd.it/q22mgl
@r_devops
What level of knowledge is needed in each area?

As someone with very broad interests who always worked best in a jack-of-all-trades role, DevOps is an incredibly exciting possibility to me. While I'm sure it would be great for a DevOps pro to be a networking SME and a security SME (etc etc), I also know that it's not realistic to specialize in everything. So I'm curious, what levels of learning in each respective area do DevOps professionals need at their entry, mid, and expert levels?

For example(s): Is CCNA all you need to get started? Should a Sr. level DevOps pro have CCIE level knowledge of networking, or is that overkill? Does Security+ give you what you need to manage, or would CySA+ and PenTest+ level knowledge be more appropriate? Would a Jr DevOps pro benefit from getting the CCNP or would things like Terraform, Ansible, Kubernetes be better to focus on at that point?

(I know that certs are not the end-all, be-all, but it's also an easier way to get the idea across. Please feel free to use specific topics or learning areas if you'd like)

​

EDIT: Mods, if this question is too broad, please let me know and I will happily pare it down.

https://redd.it/q241a9
@r_devops
Ansible CrowdStrike only run as Playbook but need role

I'm struggling and would be super appreciative for any advice. I installed CrowdStrikes Falcon sensor using a playbook and it works great.
I'm supposed to make a playbook that prompts users for entry. Their input chooses which role gets started.
I've got roles that run patches, place files, ETC.

This Falcon sensor install can't be run as a role. I googled, "make one playbook start another." This isn't advisable I understand.
Is there a somewhat respectable workaround to what I'm looking to do?

Things I've looked into:
Trying to run script quietly that starts another playbook. - The shell doesn't allow nohup & tricks.

Adding a cron job that runs a script to start a playbook. - I need the playbook to run 5 seconds from now/after the current playbook ends.

https://redd.it/q24lol
@r_devops