Life of an DevOps Engineer

1. Birth
2. Server load is too high, we need to stop all tenants.
4. Death

https://redd.it/pujupa
@r_devops

Cheap/free activity tracking application for personal use (linux)?

I do many tasks a day. Sometime so many that i forgot to report them in Jira. Years ago i used some app which gave me history of my actions during the day. For example it showed me that i had somefile.txt opened in some editor from 10:32 to 15:21. Such information would be VERY usefull to fill the jiras timesheets or any other task management app.

So i need app which:

\- tracks the apps opened and times of opening

\- works on linux

\- local app or with self-hosted server (docker) highly prefered. But i consider things like Clockify or Timecamp too ...

​

Please, only personal experience! No google results!

https://redd.it/pui5uv
@r_devops

install multiple version of node using nvm and it should be accessible by all users

I was asked to create a Dockerfile to install node v6 and node v14.16.7 using nvm and all the other non-root users able to use both node versions also they should be able to change the version of node using the nvm. I searched as much as i can but couldn't find any solution, if you have any thought on this issue please reply. TIA

https://redd.it/pummdl
@r_devops

Ask an expert: Why you want your feature flags and product analytics working together

An interview with Split solution engineer David Martin: https://mixpanel.com/blog/feature-flags-and-product-analytics-working-together/

"With modern product analytics, you know everything about your funnel, retention rates, conversions, and more. When you deploy feature flags in your funnel, you’re essentially skewing a percentage of your traffic. So while you’re on your way to uncovering bugs or user experience hiccups, now your analytics are confused, and it wouldn’t be incorrect to say you’ve created two funnels instead of one."

https://redd.it/punyhv
@r_devops

How should I take my next step?

Hey guys, I'm gonna tell a bit of history from myself:

19 months ago, I started working as a intern in a software company. 7 months later, I got hired as an Infraestructure Analyst. My responsabilites were from editing client's logos to releasing and deploying our applications. I got to know Jenkins and how CI works in general, and I enjoyed it. Since march, I have been a one-man team. The guy who worked with me and taught most things to me found a better job and left.

Since I'm a fast learner, I had no problems dealing with our CI/CD pipeline by myself so things went fine. I took the opportunity to show my supervisors that I was good and that they could trust me.

My boss said that he was going to find a replacement for the other guy. This replacement never came. Instead, I got a 40% raise. But my salary was pretty below average in the first place. What I thought was a great ideia to show my skills and proactivity, became almost like an nightmare. It is like they thought "Well, if he can manage, there is no need to hire another person".

I'm currently overworked, alone, and having to deal with a bunch of things. I'm responsible for our entire CI/CD pipeline. I have to give support for the devs who don't know how to use their credentials to connect to our TFS repository or even install Visual Studio. Or set a environment PATH. Like, they think I'm just an IT guy who is there to solve everything about their computers. I literally heard "Can you do that for me? All you have to do is to click a button". I have to edit logos for clients. I have to deal with problems that happens in some client's Infraestructure.

My boss even gave me some tasks of development which requires C# and ASP.NET MVC knowledge, which used to be my former colleague's tasks. I talked with him about a raise for the new development responsabilities and he said "I already gave you a raise".

Problem is that I made a mistake of not pursuing something in specific. What I know is a bit of everything. What I know from DevOps are the concepts of CI/CD and Jenkins from my day to day use. I want to pursue path in this field. Problem is, I'm having a really difficult time finding time to learn in depth. My job is almost non stop everyday, and I still have college in the night. I want to find a better company, but by my resume, those company won't even look at me. They don't want a bit of everything, they want a thing in specific. And I can't leave my currently job without another one in the chamber because of financial reasons.

Do you guys have some tips of what can I do, like a path so I can start going in depth to this DevOps field? It is like I finally found something that I want to work on, things that triggers my brain to solve problems and find solutions. When I automated a process that took like 4 hours every thursday to do, I felt like a superman. Never felt something like this in a job before. But I want to leave to a place where I can put that in full development, instead of where I am right now.

Thanks in advance and sorry for my english.

https://redd.it/pub9bc
@r_devops

Building a Pipeline

Five years ago I read The Phoenix Project and was inspired to dive into DevOps. As a developer I was never really satisfied just writing my code and sending it off to someone else to deploy it. After reading The Phoenix Project and tearing through the DevOps Handbook and so many other books I found that this was the wrong way of developing. So I took what I learned and started applying it to my personal projects and brought it into my workplace.

I found along the way that there were other elements from a team dynamics side that weren't often discussed and how teams can go about trying to get feedback from customers. Nor could I find a book that gave practical examples on how to take the code I was writing and show me how to deliver it to someone else other than piecing together Stackoverflow posts, random blog articles, and lots of trial and error.

What resulted was an idea for a book that brings all of this together using Go as the core language. I chose Go because it's simple to read and it's a language I use every day. I am still writing this book and the publisher provides the ability to give feedback and ask questions in their live book. It can be found here: https://mng.bz/q5P6 and their is a 50% coupon available if you use mlholmes4 by Sept 27th.

https://redd.it/puntt2
@r_devops

ideas on Managing thousands of users

We have roughly 4,000 MongoDB users across 500 different databases (mostly application users). What is the best way to manage this? The previous team imported all these users to Terraform, but I find it weird that Terraform is managing users. Then since we inherited this, we are using Terraform to provision new users on a specific database.

I am just wondering if there is a open-source that does this already that manage tons of users.

https://redd.it/pu99qg
@r_devops

Let's Encrypt Root Certificate is expiring soon!!

Hello everyone. I have some questions related to let's encrypt. If you could provide me with any hints or suggestions, it will be beneficial for me.

1. What are the differences between DST Root CA X3, ISRG Root X1 and R3 certificates?

2. Only the Root CA X3 certificate expiry date is the same for all browsers but the expiry date of other certificates is different. What is the reason?

3. Let's Encrypt is ready for expiration with its root certificate (DST Root CA X3) and will expire on September 30, 2021. If any device wants to connect with the server with the Root certificate, will it cause any problem? And if the certificate expires, will it cause any problem to the website?

Thank you.

https://redd.it/pu11xr
@r_devops

How does artifactory handle multiple concurrent downloads and uploads?

I'm trying to implement a Jenkins pipeline where we upload a new platform package when something gets committed in our SCM repo. However, multiple users rely on that package and may need to download it multiple times a day. Does artifactory handle concurrent downloads and uploads well? Or do I need to have a resource locking within our Jenkins pipelines?

https://redd.it/put99z
@r_devops

Dissapointed from salary , but patience is the key , or not ?

Hello , currently based in Germany , having round 1 year experience with Devops , mostly working with AWS, strong skills are also monitoring installation for logs or server metrics . I know also many AWS Services , good knowledge of Gitlab ci/cd. In my Company earning 25k/year yeah , 25k :( .My company its not so big and they say that i don't have many tasks to do yet (kind of a lie), that's why they can't pay me more. Many times i thought searching something else but trusting not so much to myself and my skills i don't want to risk it and suffering to find another job. I think i need some more experience including Terraform and Ansible , and some more python skills. So atm i think patience is the key and when i feel ready i can do the move. What do you think ?Thanks in purpose.

https://redd.it/puuaue
@r_devops

How to setup nginx for caching request response on a server side?

Hello.

I am looking for something that would cache the response of request like

/my/calculator?x=10&y=20&z=30

The cache should be stored somewhere on server side, and probably I could do some fancy redis here, but since this is quite a basic thing I bet nginx can do that on its own.

I want to cut off users from reaching the backend engine (php or something) since this would eat significant memory.

It would be also nice to be able to clear that cache programatically.


Edit: Okay Nginx Plus solves that with cache purging

https://redd.it/pubmc6
@r_devops

Run a script every time someone pushes to git

I'm working on automating our test process and want to build, install and run my test scripts every time one of the devs does a git push to one of the relevant projects. I'm currently using Ansible to automate everything and notify me if something isn't working but I haven't found a good solution for daily scheduling (currently using crontab but there's some limitations with Ansible's become feature) and triggering a script to run from a git push.

Ansible Tower looks like it will do most of what I need but it's outside the budget of our organization (I'm running more hosts than the community version will support)

Any suggestions for daily scheduling and triggering a build on any git push to our repository?

edit: a word

https://redd.it/pu2i4d
@r_devops

dev on-calls

I work for almost 10y in IT company (\~150 employees) as a software engineer (dev). I've seen a lot of changes regarding the product lifecycle, one was to start the on-call duty by dev almost 6 years ago. The dev on-call means that when the alarm goes off you must restore the application, root cause analysis can be conducted later. Me as a Software Engineer accepted this, I had my doubts (fear?), but after all its dev that writes the software and must be able to run it - that's how I personally rationalized this.

As the time passed I think the dev on-call is a positive thing: it creates a better programmers:

you don't write too generic/abstract code
you write operationally-cheaper software
you expose proper metrics
you naturally log proper events
whilst still being able to create good maintainable code
etc.

The on-call didn't last long, soon the dedicated support team was hired, silos were created, dev had little insight into product prod-behavior.

Fast-forward present day: I'm in different project in the same company, project started over 2y ago but gained traction last year. Since the last year, the project consists of 4 teams now and requires on-calls. I don't want to hire dedicated 'support' team, I want all (with regard to sensible team-size defaults: https://sre.google/sre-book/being-on-call/) experienced-enough devs to perform the on-call.

I made that clear from the beginning. However, currently the on-call is performed by my team only (since we're most experienced overall and our area of expertise is infrastructure/backend)

Building that team was hard, people were unwilling to take that responsibility, but somehow it took off. The team consists of devs only (like the other teams).

As the application development progressed we started to lose control over technologies introduced by each team. They create their own services, they want freedom of tech-choice and they want to do it fast. To be honest: I don't feel anymore that I should support all of that... It's also them that should feel the responsibility, I'm the same dev as they are. The point was not to be the martyr of the project, the point is to take the responsibility and create better software. Moreover people learnt that they can trade their crippled code into more operational load because it's not them supporting it

Soon I want to engage other people in on-call, the team leaders seem to understand my point of view, the devs however: do not. I hear a lot of complaining: "I don't want to do this", "I can't", "I don't know how", some people said that they will quit. Even one guy from my team wants to leave the team and demands discontinuation of on-calls in the name of "I want to focus on dev" (he had 1 incident past 6 months, he's the author of big part of our backend logic so he knows all of the details, I was shocked when he said that)

Having said all of the above:

1. Is there something inherently wrong with my goals and reasoning? I think that dev performing the on-call duty is the right thing to do (I'm a dev) however my company seems to think opposite
2. How does it work in your company? Do you have dev on-calls? Who is the first line? Do people just accept it (of course the on-call should be compensated but even though do the people accept it)?
3. If my goals are right, what should I do to create the environment where all team members feel the responsibility, understand it and are willing to participate in on-calls?

https://redd.it/ptzhx9
@r_devops

Personal domain name suggestion

I need suggestions to purchase a domain for myself. I am an entrepreneur. Someone already takes FirstLast.com, and it expires in 2022. Firstname.com is super expansive. First.net, First.org, First. Me, First.io, and others are still available. Someone also endures Last.com and honestly speaking, I don't want to use my last name as a domain name because my last name contains 11 characters itself. And my name is also five characters, so 16 characters domain looks quite weird to me. So please suggest your opinions.

https://redd.it/puzesf
@r_devops

Career change, trying to from QA to DevOps / Cloud?

Hi group,

I'm trying to switch careers here. I have over 4.5 years of QA experience with automation and I know java so programming is not a problem. I was fed up with my job and thought of switching job. Embarked on a Business Analysis journey, got a post graduate certificate in it and got a job at a small organisation. Not liking the job at all. I'm thinking of switching to SRE roles or DevOps / cloud . As there seems to be good earning there as compared to BA.

Would appreciate if you could lay down the roadmap. Where do I get started and what is the realistic expectation salary and job prospects for someone starting out fresh.

Thank you

https://redd.it/puxwiy
@r_devops

How long to give a new job?

Hey guys, just looking for some advice from anyone who might have been in a similar situation..

Basically, I worked as as DevOps Engineer for a fortune 50 company for the past few years, usual big slow corporate stuff with highly restrictive RBAC so I only had to worry about my immediate project's services and CI/CD and we worked in a pretty good scrum system that always gave us a fair workload. It was a cushy number but the pandemic caused me some pretty bad burnout so I decided it was time for a change, I got a new job a few months ago and was hoping a fresh start would give me a mental kickstart.

So far, I'm more stressed than I've been since the first few months into my first job and things don't seem to be getting much better. I'm working for a much smaller consultancy and have a vast amount of stuff I need to learn, at the same time I got thrown a major project in my second week that was already behind schedule and its been a bit of a car crash from my POV, I've had to learn everything on the fly while at the same time supposedly acting as an SME to the client. I just figured out at 5pm this evening that the entire pipeline I'm currently working on is a non-runner because of technical limitations, I've wasted two weeks of effort on nothing so Monday morning is gonna be fun explaining to the client!

In my previous company that would have been flagged in a spike but here it seems I just get assigned these huge complex tasks with tight deadline and if there's an issue half way through its my problem.

TLDR, out of my depth at new job, work systems are a bit chaotic with no clear planning or acceptance criteria and because its such a small company there's no respite and nobody to fall back on.

I've never left a job after such a short amount of time but definitely considering it at the moment, just wondering if anyone has done similar, any advice on how to navigate the shitstorm or low long I should give it before making a decision? Leaving a job after a few months is presumably a red flag for recruiters but I don't want to throw away a year of my life being miserable just for the sake of some work experience.

https://redd.it/puuw3t
@r_devops

Not able to run docker container in Mac M1. Help please!

Hi,
I'm a front end developer, recently I was given a take home test to do. The problem is they have included the BE code and docker in the project too. I'm supposed to be running docker locally with nginx. They have set it all up with docker-compose file.
Only problem is I'm not able run couple of containers including FE one. It's showing an error and exiting immediately. They did say I'm not supposed to publish test to public or else I'll be disqualified. So I'm in bit of pickle. Hoping one of you guys maybe able to help me?
The error is
runtime: failed to create new OS thread (have 2 already; errno=22) fatal error: newosproc

I did Google this issue and I saw couple of stack over flow answers but I didn't understand what I'm supposed to do.
Thanks for reading 🙂

https://redd.it/pv2k2j
@r_devops

How to mix separated versions of Python in the cleanest way

Hello

My service is running 120+ GHE repositories with different versions of python engines - each of them requires a separated venv.

On my previous laptop running Ubuntu LTS 18.04, I was fighting with python versions (default-system is python 2.9, and no official support after Python 3.6 from canonical repositories).

How do you handle properly, multiple versions of python engines, with large amount of venvs (pipenv or poetry based) ? I just have migrated to Ubuntu LTS 20.04.

Thank you

https://redd.it/pv41m8
@r_devops

What is the best Shared filesystem solution for Kubernetes cluster (AWS EKS) ?



We have an EKS cluster on three availability zones, our pods need a shared filesystem across these availability zones so they can share what they write on it, we started by implementing a native solution in AWS which is EFS (Elastic Filesystem) but since this solution is based on NFS, it caused significant latency on the filesystem, so we tried to use CephFS implemented with Rook, this solution worked very well with few pods writing to the disk but then when the number of operations increased we could see that the latency increased as well.

My question is, do we have an alternative solution to install on our cluster that can be scalable (Provides the same Throughput independently from the number of operations on the filesystem)?

https://redd.it/pv5rtz
@r_devops

hashicorp consul understanding

Hi All,

We are looking into consul for service discovery, and wondering.. do user application actually connect to the consul servers for resolution of the apps ?

Or do you guys purely use the service discovery for the internal workings of your microservices? (as my understanding is you need a consul agent to get the dns resolution.. but general user machines would not have the consul agent) So do you use other methods of resolution for the user applications for where the services are located?

For example user-application-a needs to resolve service-z, does user-application-a connect to consul in some fashion, or perhaps user-application-a connects to a specific database that is populated with dns records from consul?

thank you!

https://redd.it/ptz2mk
@r_devops

GitLab CI/CD example project?

Looking for an end-to-end CI/CD example project based on GitLab. Any help, please?

https://redd.it/pv98oo
@r_devops