CREATING BASE DOCKER IMAGES WITH ALL DEPENDENCIES INSTALLED

We've got a bunch of microservices that we build via docker. Right now everytime we build an image we keep installing external dependencies/packages.

I'm thinking of creating a base image which will already have all of our dependencies/packages installed and then on top of that image we just build our microservice.

Do you folks recommend this kind of approach?


Edit : Sorry if I look like I'm yelling in the title, I couldn't do lowercase for some reason

https://redd.it/12srdou
@r_devops

When applying for roles, how much does the purpose/vision of it affect your decision?

I’ve been selective in trying to land my first engineering role because the visions haven’t seemed super appealing to me.

To you, does the product/purpose make a difference? Or do you just pick roles based on skills you’ll build and the technology you’ll be using?

https://redd.it/12t2kah
@r_devops

Noob here looking for a method to install Docker Desktop edition to a different partition in Win 10.

Looking for a foolproof method to install Docker Desktop edition to a different partition in Win 10. I tried the following command line flag but it didn't work. Installation proceeded with default settings i.e on the C drive itself which by the way is not having much free space at the moment.


* --installation-dir=<path> : changes the default installation location (C:\\Program Files\\Docker\\Docker)

Would be grateful for any pointers to accomplish this.

Thanks

https://redd.it/12t2gpk
@r_devops

Hi guys!!! I new to CICD pipelines. Can anyone recommend any instructor/YouTube or resources I can use to master GitHub Actions & Jenkins.

YouTube channels, book recommendations etc will be highly appreciated.

https://redd.it/12t236w
@r_devops

Small environment on docker swarm

Hey devops!

I recently jumped into the world of devops with docker swarm. After a somehow difficult learning curve (mostly because of the environment constraints for the swarm setup) I am thrilled and loving it.

The thing is, I took over the environment mid project and quite frankly is a bit of a clusterf***. But that is another story.

I need a quick, easy to setup and if possible web hi enabled configuration management system to at least fix the mess and create a platform basis for something better (i.e ansible, Salt, the likes).

In essence I have to modify a bunch of parameters on more or less 200 Debian servers (apt sources, create sudoers, ssh keys, etc). I have ssh access to them all.

I thought initially on a dockerized ansible but there seems not to be an up to date ansible container and to be honest I don’t have time to create my own, as there are multiple fires to deal with. I need something quick, easy and functional.

I know I may be looking for the golden triangle (easy, beautiful, quick) but….is there such a thing?

https://redd.it/12tctuw
@r_devops

Summer internship

How I can find a summer internship in DevOps

https://redd.it/12tgiuo
@r_devops

Terraform Cloud Discord Webhook Proxy golang app

I made with the help of ChatGPT a proxy that forwards the webhook from Terraform Cloud to a Discord Server to have the notification there.

https://github.com/smark91/terraform-cloud-discord-webhook-proxy

Let me know what you think and send suggestions. I hope you like it!

https://redd.it/12tkjhb
@r_devops

How can I learn puppet with only one node?

Hi, I'm trying to learn puppet but all the tutorials and quickstart guides assume you have access to multiple nodes. I only have one. Is there a way to run puppet in a single node mode?

https://redd.it/12tm3t6
@r_devops

Database / Platform to track releases done on each system - Does it exist?

At least in my environments, because of the large amount of microservices / apps we now need a central platform track all the versions released on each system.

I am imagining if one was try to build something DIY, it would be some sort of centralized database or even simpler - an excel sheet with every version and system listed.

I can't seem to find a known tool that will do this (OSS or paid). Does anyone know what is the proper term for such a platform or what are the keywords / terms should be used.

Supposively Spotify Backstage can do this sort of release tracking but that is with the assumption that app was started / onboarded from Backstage initially.

https://redd.it/12tmczj
@r_devops

IaC - best practices

Hello, my organization runs an application with google cloud serving as the underlying infrastructure. We manage this through a terraform infra as code script.

We now have a need to host an additional
Application with a second app code base.

Question: should I have 2 different code bases for the underlying infrastructure or a single code base?

If I elect single code base, how hard would be to decouple the code if I choose to migrate one of the apps?

Note: I will also have different code bases for the various environments (dev, test, prod) but this is irrelevant.

https://redd.it/12tmzln
@r_devops

How do you decouple metrics generation from structured logging?

Where I work, I often see monitoring code that records the same fact twice: one for logging, and one for metrics, for example (Python-ish code):

base_logger = get_logger()
REQUESTS_RECEIVED = prometheus_client.Counter('requests_received', 'Number of requests received', ['method', 'status'])

def middleware(request, next_call):
logger = base_logger.bind(method=request.iss.onethod, endpoint=request.url)
logger.debug("starting to handle request")
response = next_call(request)
logger.debug("successfully handled request", status=response.status)
REQUESTS_RECEIVED.labels(request.iss.onethod, response.status).inc()

Having thought about it a little, it seems that it'd be better for the application to export only structured events. Some downstream process then receives these events, processes them, and then generates metrics based on defined rules. For example, one could define a rule for the above `REQUESTS_RECEIVED` metric, and another for a `REQUEST_TIME` metric based on how long it took for a request to be handled. In particular, those metrics could be derived from multiple events in aggregate.

Does something exist already to do this?

https://redd.it/12tuwyk
@r_devops

How to create a multi developer bitbucket workflow

Hi guys,

I have started at an agency who has multiple developers all over the world.

They have been all cloning sites from the repo to their local machines, making changes and committing. But you know there are a lot of problems with this. People forget to rebase to main. Developers are all using different versions of Apache and mysql not to mention php.

And then they commit to main and FTP to staging thinking “oh it worked in local”, something breaks and hopefully they notice before pushing to production.

So.

This is the process I am setting up:

1. Using source tree on local to commit to bitbucket.
2. Bitbucket pipeline runs automatically to staging.
3. Bitbucket pipeline is configured to manually deploy to production.

But the missing pieces of the puzzle are still different Apache versions, MySQL and php. So

Is there a software that a team of devs can use that will automatically set the versions of AMP so when they commit we are getting consistency from each developer?

Or is there a way to require the versions in the Git config file.

Thanks very much 😊

https://redd.it/12tvrcg
@r_devops

what is needed to be an SRE/Devops/platform in MAANG

Hey All,


I think this will be usefull for many

Any one who works as SRE/Devops/platform anything that they call these days in MAANG
what is the necessary skill needed to get inside these companies as SRE/Devops/platform engineer


I have a total of 6.5 years of experience in various cloud and devops tool including terraform , ansible, packer k8s docker CI/CD jenkins/ansible/team city . I mostly work in central team so i get requirement to automate CI/CD end to end so i use mix of shell /python /powershell and mix of tools


Now i want to move higher up in companies , I am a decent coder but have never done web dev something of that sort , mostly mine was scripting . I am good in fixing systems ie something breaks , i can dig deeper like dns /lb something got broke i will get the first call to fix since good at linux/networking and troubleshooting


I heard MAANG kind of companies just evaluate problem solving and never care about experience is it true , if so then how should i start and what all i need to revise please help


and how should i prepare now for these roles in these companies ?

https://redd.it/12twdrf
@r_devops

Bit of a "wow" this company is great to work for post, feel free to share stories yourselves

So i started at a big financial tech firm recently, Ive had my share of shitty jobs in the past where blame culture is massive, but I was shocked a few weeks ago by this new company.


Someone in another team fat fingered an update in a tool which took down everything in production, in such a away that the automated systems couldn't bring the apps back up successfully, so a bit of manual work had to be done, it was over a weekend so it luckily wasn't a lot of traffic hitting the app. But it was still a big outage.


In other places I've worked even though It was accidental with no malice intent, that guy would be fired immediately and the issues swept under a rug(fixed quietly) .


In this company they didn't name names, in a company wide email, they said it was a fault in an automated system and if anyone was to blame it would have been them as managers for not foreseeing this would be a problem (email sent by ceo, cto, coo, etc)

Only a few people in cloud ops with the correct access to this tool could see who did it.

The guy later said he was called into a meeting with all the ceo, coo, types, him thinking the worst.

They reassured him the blame was on them, thanked him for alerting them to the issue, and seeing how distraught he was offered the next few day off paid so he could return to work in a good frame of mind.

The guy is still at his job and is part of the team fixing the issue, still has all his admin accesses too.

https://redd.it/12txhbs
@r_devops

How DevOps and QA can be friends

Fluid 🌊
Flexible ✍🏻
And fun? 😁

Join us for an epic conversation with Doug Simons - the inventor of SenseTalk - to discover the hottest trends in #lowcode #softwaretesting and how they empower your dev and QA to work together more efficiently.


Register now: ow.ly/sBQN50NGXMx

https://redd.it/12tuvxy
@r_devops

Looking for the cheapest deployment option

Hey guys. I need to deploy a service into AWS and I'm divided between ECS Fargate and Lambda functions. It's a simple service which gets e-commerce orders and transfer to a database, parsing the information. It will probably be called between 10 to 50 times a day.
I already have the container ready to be deployed but I'm also thinking about lambda functions due to its low price... What do you guys think? The service itself takes just a few 'ms' to run.

https://redd.it/12u1xt6
@r_devops

Anybody adopted OpenTelemetry for all observability signals (logs, metrics, and traces)? If so, any thoughts?

Looks like there have been a lot of recent advancements with OTel, and I know traces were always had first-class support. I'm curious if anybody has also adopted OTel to handle their metrics and logs as well. If so, what's the good/bad/ugly?

https://redd.it/12u3g7z
@r_devops

How do you deploy secret files within a pipeline?

I currently have a bitbucket pipeline that deploys to docker hub a image and thats fine and all but now i need a way to also include a file (a .crt certificate) while building the image and deploying it to docker hub.

I am imagining holding said file in a bucket, something called PRODUCT-secrets, then copying all contents of the bucket to a folder within the image, but that would kinda suck to do.

The other alternative of course if to simply commit the file in the source code but then every developer would have access to that file, making it a security risk.

There is also the other alternative which is to have the files in the VM that would have the image be deployed onto and bind mount them into the container, but then when the VM gets destroyed the files would be lost, kinda bad as well.

&#x200B;

Is there a better way to do this?

https://redd.it/12u5fpb
@r_devops

Broken websites: are they 2x as frustrating because we know how to fix them?

Specifically: 404 on https://careers.etsy.com/ca/

Or broken profile views when an ad blocker is installed: https://www.reddit.com/r/Etsy/comments/11512uh/favorites\_not\_loading\_properly/

Sometimes it seems like reddit or some other "social media" site is the only way to report these things.

Is there an alternative? Like a "report problem" link with capcha. Who does this well?

Or hell, just checking your links with a scanner or checking your logs for errors.

https://redd.it/12udo4x
@r_devops

How Useful has Distributed Tracing Been for You?

Do you think it’s worth investing in? What pain points do you have?

https://redd.it/12ufbqh
@r_devops

Five Rookie Mistakes with Kubernetes on AWS

We recently setup a new Kubernetes cluster on AWS and made some rookie mistakes. In a nutshell:

Not setting resource memory limits
Using EBS storage with more then one AZ
Using the default instance templates
Not using version control for your config

Here is the full story Five Rookie Mistakes with Kubernetes on AWS

Can you think of any others or any that you have stumbled over yourself?

https://redd.it/12ugpqp
@r_devops