Differences between implementing Centralized Logging vs Distributed Logging in a small scale IT industry having applications running on AWS?

Which one would you chose if you care more about monitoring your infrastructure, specially if you have a small team and limited amount of resources? I am new to DevOps culture.

https://redd.it/dvp69a
@r_devops

I fell into this field and I'm not successful at it. The operations people do not understand how I work as a software engineer.

There is a pattern wherein I get tasked with a difficult, large project that nobody wants to touch. Sometimes after spending a great deal of time figuring out what not to do or finding something that doesn't work, I stake out a robust v1.0 project in something I've never done before that solves the problem and I deliver on it. I generally approach other issues solving the real question or dilemma, with a lot of deliberation and caution because it's production.

This seems unwelcome in a "move fast and break things" culture that is all too prevalent.

In executing the project, I get whatever helping of the following: zero support from management or coworkers because everyone is swamped and ignoring emails and chats, my boss doesn't understand what I'm doing or have the time to care, his boss doesn't have a clear idea because there's no communication between them, or the absolute worst, management with no authority to deal with other teams.

I have to rely on my intuition to navigate complex ecosystems made difficult by the fact that everyone relies on dated, legacy software and OSs. I find this intensive, isolated trial-and-error process to be immensely frustrating and taxing, so to burn off steam here and there and get my mind off it so I can go back with a fresh set of eyes, I might check my phone or look at something else online for whatever period of time which pisses off management despite the fact that people are *always* screwing around here. Anyways, this informs them that my performance is not up to par, usually something I usually find out too late in the process.

I cannot use 100% of my brain all the time to do this stuff. I don't know how anyone can.

The operations people especially do not understand me. They want MVPs that don't actually work (we would have been shut down if the MVP they were asking for leaked confidential information, which it would have), this last company had no code review and had things constantly breaking, and my boss's boss's deploy scripts had the worst code I've ever seen. I cannot, with my experience, move at the speed they're asking because what they're asking for doesn't fucking work.

Moreover, the operations people need tremendous reliability in employees so it's not accommodating for health issues despite what they think. As an epileptic amongst a handful of other issues I've been disciplined twice for having seizures at night that severely interfere with my work the next day and I have to call off late at night because I didn't get any sleep. Everyone sick has to WFH. I don't get it. I get that these are "small teams" but WTF, fix your hiring culture and make them larger.

Hiring in devops is another can of worms. This field is notoriously picky and has swamped engineers as a result.

I'm no longer calling myself devops but what I do best: software and infrastructure engineering. With this knowledge I'm looking for my next gig and I hope I find it fast. I've learned these lessons too late.

https://redd.it/dvo7zj
@r_devops

Is there any downside to using Bitnami images from AWS marketplace?

I see that many images are available free to use whereas Enterprise versions would be expensive to pay directly (looking at jFrog artifactory).

However it sounds too good to be true that you can get all these Enterprise-grade OSS from a third party, with certified images in the AWS marketplace, without paying any money to original vendor (jFrog) or Bitnami.

https://redd.it/dxdouf
@r_devops

Docker, GitHub and continuous deployment

Hello guys,

I recently moved away from Heroku and created a droplet on Digital Ocean. Now with my new won freedom I also lost some luxury functions like the auto deployment functionality of Heroku but I bet we can figure out a better solution for this :^)

Well first of all I am wondering how my approach should go down. I have a nodejs backend which I want to deploy to my VPS when I commit to the master branch. I have a docker-compose file for the database right now and I plan on dockerizing the backend as well so I can add that to the docker-compose file.

So I am wondering what my plan of attack should be? I will create a Dockerfile for the nodejs backend. So when I push to master, I need to get the whole repo to my VPS, create the image of the backend and launch the docker-compose. Any recommendations? I do not want to install an additional Jenkins on my VPS. In the past I used Gitlabs CI runner but I am on GitHub right now.
Side note: Where do I save my environment variables? If I put them in the Dockerfile, I would have to check them into git as well

https://redd.it/dxbacd
@r_devops

Is it possible to come from a non-dev job into a DevOps career?

First off, I understand that DevOps isn't meant to be an Engineer role. It's meant to be a culture. Now that we've addressed the elephant in the room, we can move onto the reality of the fact that DevOps Engineer is now a real career title in the workforce.

I've read The Phoenix Project. I have done some studying and playing with tools/technologies (to list them in a mixed-hat fashion) such as Ansible, Terraform, AWS, GCP, K8s, OpenShift, GitLab CI, and Docker, and trying to utilize these tools at work where applicable. I'm also kind of basically dangerous at writing Helm charts.

For my work I have been writing useful web applications in Golang, and pushing them to GitLab where it runs my CI for unit testing and then creating a container image. The CI then pushes the new image to an OpenShift project image stream, which triggers my deployment to start a rolling update for the new code. These tools get used by people in my department, as well as some remote teams.

Other than that, I have submitted snippets of Ansible playbooks to our team that's in charge of that infrastructure automation to help prevent future outages after we deemed that the outage could have been prevented by an OS configuration.

My biggest take away from job interviews thus far has been that my experience is inadequate, which I believe is partially due to the fact that my actual title is more akin to Help Desk Technician II. Another possibility being that I don't have a college degree. I haven't been able to use Terraform, AWS, or GCP really at work because we host our own servers. What more should I be doing to advance in my career? Is there a stepping stone position from where I'm currently at that people tend to go through first? I'm partially asking all this for real, but I suppose also partially to vent after many promising interviews ended with no call-back.

https://redd.it/dx7gux
@r_devops

Looking for an AWS centered Airflow/job/batch management (python)

Hi,

I'm having troubles finding projects like airflow/luigi but that would be centered on AWS.

My goal is to :

* run anychronous but still dependent processes in ECS
* write workflows/tasks as code
* uses cloudwatch for log storage and retrieval
* uses cloudwatch for metrics storage, retrival being done with grafana
* be able to use aa REST api to launch the workflows, monitor them, etc...

The absolute requirements:

* free
* open source, hackable
* made in python

I checked the airflow documentation, it seems it'd a great product but it's doing too much things by itself, that could instead use actual AWS services. It has some integrations with lambdas and S3 but it seems like afterthoughts, even if they may work properly.

How do you guys handle this ? I don't want to spend my time rewriting dashboards that would analyze cloudwatch, ECS logs, etc..

Thanks !

https://redd.it/dx6vo9
@r_devops

When do you decide that you need to use containerization in production? what is the benefits?do you have a checklist or something?



https://redd.it/dx3864
@r_devops

Kibana is not responding.

Kibana is not responding.

Please see this https://pastebin.com/LW0QY1Ma

what is the issue ?

https://redd.it/dvpzgj
@r_devops

Need Help Starting Out

Hey guys,

I think I might be in over my head here. I wanted to get a custom IDE for my startup built. So we found Eclipse Che and figured it would be a good starting point to build on. I am not a technical person at all and just looking at the landscape of it out is overwhelming. I would love to have all the cool features of Gitlab and I think it is possible with all the extensions and plugins Che has floating about. I guess my question is how do I get everything I need done without being a technical person? Where do I find an Eclipse expert?

https://redd.it/dvlfl5
@r_devops

Why do you still use Jenkins?


I mean seriously it is an absolute dinosaur. Hard to manage, difficult to scale. Why do so many still use it over the more modern alternatives like Circle CI or Gitlab? Am I missing something here?

https://redd.it/dv81r1
@r_devops

New to this, probably not doing things correctly. Jenkins/Artifactory

I work for a large company with an existing infrastructure that I'm trying to piggy-back off of. I maintain mostly SQL scripts and other non-compile-ables in a Bitbucket repo. I'm trying to establish a consistent push from Bitbucket to a Windows server (overwriting existing files).

Initially, I used Jenkins (on Linux) mounted the Windows directory and cp'd the files from the workspace to the Win server. Basic but it worked, though it required sudo to mount the remote path.

We're supposed to push any Jenkins build artifacts into Artifactory though, and I shouldn't be using elevated permissions. So I've changed changed things up to deposit build artifacts to Artifactory and SSH a zip of the files to my win server, I can use a ps1 script to extract where they need to go - but I have to do that manually or scheduled (is there a way to trigger a remote ps1 script or a SQL Agent job from Jenkins?).

I really don't know Artifactory at all, and not sure if I should be trying to use Jenkins to push from Artifactory to my server, or pulling Artifactory artifacts from the Windows server on a schedule, or deploying from Jenkins to both Artifactory and Windows separately.

I also am utilizing SonarQube via Jenkins, but that's working fine currently.

**TL;DR:** What is the best way to move scripts from Bitbucket to a Windows server via Jenkins (on Ubuntu)?

https://redd.it/dvkw90
@r_devops

Composition over inheritance applied to Docker

The main benefit of Docker containers is that they are self-contained. For developers, that means one just needs to inherit from the desired Docker image that contains the necessary required dependencies, and presto, one can build one’s application deliver it to production. Most of the times, the process is pretty straightforward. Containerization allows scaling on unprecedented scale.

However, the downside of this self-containedness is that updating the parent image(s) becomes a nightmare.

https://blog.frankel.ch/composition-over-inheritance-applied-docker/

https://redd.it/dxo4sa
@r_devops

Issue with login to ECR inside Jenkins build agent

I am using Jenkins to build something and I want it to run a custom image that is hosted on ECR. The keys have admin access (testing only - will restrict later) when I go to put it in I get a 401 error seen below. I have also included how I have my build agent set in my pipeline. I was hoping someone solved this already, google has just turned up the 'I figured it out thanks every' posts.

​

pipeline agent

agent {
docker {
image '<id>.dkr.ecr.us-east-1.amazonaws.com/node_build:latest'
registryUrl 'https://<id>.dkr.ecr.us-east-1.amazonaws.com/node_build'
registryCredentialsId 'aws-creds'
}
}

&#x200B;

Error message seen in Jenkins

Running on Jenkins in /var/lib/jenkins/workspace/Deploy UI
[Pipeline] {
[Pipeline] withEnv
[Pipeline] {
[Pipeline] withDockerRegistry
$ docker login -u <Access key> -p ******** https://<id>.dkr.ecr.us-east-1.amazonaws.com/node_build
WARNING! Using --password via the CLI is insecure. Use --password-stdin.
Error response from daemon: login attempt to https://<id>.dkr.ecr.us-east-1.amazonaws.com/v2/ failed with status: 401 Unauthorized

https://redd.it/dvh4rk
@r_devops

Testing Packer Images

Is there a better way to test Packer builds beyond the fact that there were no errors in the build process? It doesn't look like Packer has anything for hooking some kind of validation of the system before finishing the build, so I'm not sure how or where it would happen. Has anyone ever done something like this?

https://redd.it/dxoelb
@r_devops

Are there new questions/tests you're hearing in interviews you weren't getting a year ago?

Hey, I'm pretty early in my devops career and have just been getting thrashed in my last couple interviews. Most of the 'here are questions you need to prep for' articles are a year old at best and I've read a ton of them.

Any newer ones you might be asking your candidates or are hearing yourselves that you didn't have a year ago?

https://redd.it/dxo6pt
@r_devops

Lambda function not being triggered by SNS notification

I have an application that does some processing which puts out a SNS notification depending on the results. The application works fine - I can see the console log of the notification being sent and see in cloudwatch that none are failing - however I am unable to get a lambda to run once the notification is sent out. There is an existing subscription to the SNS from the Lambda which was created using the serverless framework and I have confirmed in the console that a subscription exists.

&#x200B;

What are some steps I can take to troublesoot this? I'm a bit lost today. Also any ideas on cause would be great

https://redd.it/dxq0ce
@r_devops

protobox - gRPC dependency management

We needed a way to manage the protobuf dependencies in our projects were i work, so we made a tool for it. Check it out and give it a star if it tickles your fancy.

[https://github.com/UNIwise/protobox](https://github.com/UNIwise/protobox)

https://redd.it/dxmggt
@r_devops

How do you guys manage secrets and API codes in github ?

Are you using any specific tool such as vault or console or perhaps base64 to do it?
What do you think best way to do this ?

https://redd.it/dxksw8
@r_devops

Video-game programmer (c#) getting tired of my dayjob looking for the right path

I work as a bartender part-time as I pursue making my dream projects. Bartending has been fun/lucrative for a few years but I'm getting really tired of people. My friend recommended I get AWS certified and see where that takes me. My ultimate desire is to find a half-decent paying job that is a bit more cerebral than bartending and still allows me enough time to program. Any help is very appreciated.

https://redd.it/dxvqps
@r_devops

Unrecognized services in Puppet Enterprise

I have tried to run the following:

puppet resource service puppetdb ensure=running

and receive the error

Error: Could not start Service\[puppetdb\]: Execution of '/sbin/service puppet start' returned 1: puppetdb: unrecognized service

service { 'puppetdb': ensure => 'stopped', }

I am also running into the same issue with puppetserver. I suspect this may be a certification issue, but wanted to run it by others.

https://redd.it/dxxpsd
@r_devops