Branching strategy for Infrastructure as a code

Hi folks,

I am very curious about people's way to manage infrastructure in a git and looking for a way to improve mine.
So, we are using terraform and store code separately from source code. The infrastructure code is relatively complex and on general divided by 2 modules(global resources, regional resources) then each environment call modules from own environment folder and supplying variables.
I have aproblem for branching strategy for module (it's a separate repo). The software release happens not often (1 in few month, because before going to prod it is tested on like 4 environments).
So we have a master branch, where the prod deployed from SW released version (say 2.6.30).
Then we have a develop branch with the version that reflect current development (let's say, 3.x)
Then sometimes developers start a complete different version (4.x) while releasing fixes and features for the current master and develop branches.
So,we have a sort of branching hell.
Do you guys have any tips on how to support this scheme?

https://redd.it/syi0ye
@r_devops

Which source code management alerts are most important to monitor?

Assume the alerts are sent to a central channel and not to the individual user who performs the action.

View Poll

https://redd.it/syhdy5
@r_devops

Facing an issue with Step Function

I want to run a function every 2 mins but I want to start those 2 mins only after the excution of that function is complete. I saw that this is a usecase of Step Function so when I created my workflow.

Start -> Run ECS Task -> wait(2mins) and now I want to run the function again in an infinite loop. But step funtion is not allowing that, it need an "End" statement.

Any workaround for this?

https://redd.it/syk9k2
@r_devops

Has git conquered the source control world?

Are there people who still use other source control like ClearCase, SVN, etc, other than old projects that are sort of abandoned?

https://redd.it/syo6tm
@r_devops

CRLF vs LF Git FIX

git config --global core.autocrlf true

Does this allow mac/windows users to treat CRLF and LF as the same or will this ONLY remove the warning error?
Currently when i do file comparison via filezilla it shows my files as different which is a pain when working in a team.


If I pull from github when synced with server it shows conflicts.
I think this is due to me being windows and team being Mac so I'm looking for a long term solution!

https://redd.it/sylkgl
@r_devops

DevOps projects

How are you guys doing of presenting DevOps projects in your resumes.
How to do or demonstrate a DevOps project in GitHub?

Shoot your opinions and answers

#devops

https://redd.it/syk3ae
@r_devops

Remote work cost of living calculator

There’s been a few discussions on here recently about remote salaries and how they normally vary by location based on cost of living.

Is there any online resource we can use to predict what companies cost of living based salaries will be?

Example, I’m currently looking to move out of NYC area but don’t want to totally screw up my salary. I’m looking at Frederick Md or Dutchess County Ny. I think Maryland is safe but I’d prefer Dutchess county (Poughkeepsie area. Even tho it’s high tax and reasonably pricy I think it will fail the COL calculator.

https://redd.it/sypi9i
@r_devops

Why can't I create a standard public IP in Azure?

Hey,

I'm trying to create a Standard Public IP in Azure but when I do it errors out saying I have insufficient quota for Basic Public IPs. I know I do, that's why I'm trying to create a Standard, not a Basic because I have plenty of Standards left to use.

I originally tried this in Terraform but not I'm just doing it in the Azure portal to troubleshoot. I'm selecting Standard in the SKU so I don't know what the problem is or why it thinks I'm trying to create a Basic.

Any help would be appreciated.

Thanks


UPDATE: I deleted a Basic IP and tried to create a Standard. It worked but now when I check my quota both the Basic and Standard have gone up by 1. Does a Standard Public IP also use up your Basic Public IP quota?

https://redd.it/syks00
@r_devops

Some of the best terminal utilities you have ever used and are still using.

When I got introduced to tmux or lazygit, I got hooked to the wide range of powers 😅 that i felt within my hands. Across my org, was successfully able to move everyone to be using either of them.


So what are some other awesome terminal utilities that you have come across?

https://redd.it/syrl3g
@r_devops

Kubernetes and Cloud Native Associate (KCNA) Certification. Good for someone looking to break into the field?

I am looking to break into the DevOps field in hopes to get a Junior / Beginner level role after working in IT Support and Application/Software Support for 3-4 years.

I stumbled upon the KNCA Kubernetes and Cloud Native Associate Certification from The Linux Foundation and really liked how it’s geared towards beginners. I have been doing a lot of research and searching at job postings and found that Kubernetes is a in-demand skill.

Do you think this certification is worth while? Obviously it’s beginner based and nothing compared to something like the CKA or CKAD certifications… but was wondering if it’s a good talking point on a resume or to get some looks from recruiters/interviewers/. I was thinking of doing the certification and having some projects that use Kubernetes hosted on a GitHub or talking points where I’ve had hands on experience to supplement the learning.

For a Junior/beginner… would you recommend this? The price isn’t an issue for me at the moment. Here is the link for more info:

https://training.linuxfoundation.org/certification/kubernetes-cloud-native-associate/#review_module

https://redd.it/sz28zt
@r_devops

Does anyone know what happened to slack on 2-22-2022 ?

Hey guys, wanted to understand the cause for slack downtime.

https://redd.it/sz9ldm
@r_devops

ProxySQL cluster behavior question regarding max connections (and advantages/experience with clusting)

I'm considering running a proxysql cluster inside kubernetes and be centralized instead of running sidecars in each app's pod so I have more control over the connections.

I want to be able to make sure there's a maximum amount of connections to my db's per application regardless of how many instances of the application (pods) k8s autoscales to. Lets say the db has a quota of 10k connections. If I run proxysql sidecars, whatever max connection number I set scales as the app scales up and down.

If you're running a ProxySQL cluster and have max\_connections for a hostgroup set to 1000, will that be 1000 total or 1000 per node in the cluster?

I can't find docs that describe its behavior well but it kind of seems like the only advantage of clustering is to apply a config to 1 node and have it propagate automatically.

Does anyone have experience with doing something like this?

https://redd.it/szaw8l
@r_devops

Is it tough to put security into DevOps?

I have recently read an article that made me curious about what others think of putting security into DevOps.

Eurostar Article Link: https://huddle.eurostarsoftwaretesting.com/putting-security-into-devops-is-tougher-than-it-looks/

View Poll

https://redd.it/sz9s46
@r_devops

What are the scenarios where Jenkins is more appropriate over a bash script ?

I have been asked to prepare to take lead of more team members and responsibilities would include decisions on tools.

Have skimmed through the Jenkins reference including Pipeline and syntax like stages/steps etc. In my earlier responsibilities around 7-8 years ago, I have used bash scripts to automate build processes. I do not want to use tools just because they are industry standard/everyone is doing it, without actually understanding what problem it solves. The only thing that I could think of Jenkins solving is OS portability of the build process. This in itself isn't a priority for us. We are on MERN.

https://redd.it/szewp3
@r_devops

From this morning's standup

A colleague, describing a meeting with customers: "They were trying to nail us down on a completion date, and we were trying to nail them down on the requirements."

I can imagine the dialogue in that meeting:

"When do you think you will be done?"

"Well, that depends. What exactly do you want us to do?"

https://redd.it/szjryt
@r_devops

Grafana vs New Relic

Hello, I'm a Prometheus + Grafana user and most of my system metrics are centralized at Grafana. I'm looking to hear more opinions about New Relic users to see if it's worth it.

Some doubts I have:

1. What New Relic does better than Grafana + Prometheus?
2. How complex it is to implement a monitoring system in New Relic?
3. What New Relic can do that Grafana + Prometheus can't?

Feel free to add personal opinions about this topic too. Thanks.

https://redd.it/szl59g
@r_devops

Chef Server OSS

I've been looking and trying to understand which versions of Chef Server are open sourced. From their downloads page I sometimes get the Apache-2.0 license information for versions <=12.14.0, but not all the time. I'm having trouble finding definitive information. Is anyone else running the FOSS version of Chef and can point me in the right direction?

Do I have to build it myself instead of downloading from their 'downloads' page if I wanted it for use with a company? Does anyone have ties to the community Slack page and can send me an invite (seems locked down and only by invitation now)?

https://redd.it/szkiiu
@r_devops

How to monitor Prometheus itself

I currently use Prometheus and Alertmanager in a Kubernetes cluster for monitoring. However, if Prometheus itself or Alertmanager go down then this could only be verified.

Any ideas of another system that could be used to monitor these so there is redundancy? I'm looking for something that could integrate with Slack to send messages if Prometheus went down.

https://redd.it/szllrw
@r_devops

Debugging Azure Devops

I'm trying to build my first Azure CI/CD pipeline and to be honest I'm having a nightmare. The pipeline I'm building is just the standard tutorial one to create a storage account in Azure using an ARM template.

I keep getting the error message:

"Unexpected character encountered when parsing template"

I've checked and rechecked my ARM template and my YAML. I use the ARM add on in VS code to check the json and it seems fine. I can also manually deploy the ARM file in Azure and it works no problem.

I'm not really asking for help on the specific error, what I actually need is a good place to go to find out exactly what the error is telling me. I can't find any record of this error in MS docs and the usual places are no help. So any Azure Devops people out there, do you have a go to resource somewhere out there to help you fix problems with pipelines?

https://redd.it/szp010
@r_devops

Looking for an OPEN SOURCE/CROSS-PLATFORM deploy script. What is everyone using now a days?

I'm looking for a scripting language that I can create an installer with. I was wondering what was popular now a days. Something I can run from the linux or windows command line that will unzip packages in the right locations and will read environment variables from a text file and inject those into the appropriate config files, etc.

I'm open to automation with Terraform, Ansible, etc, is powershell my only option or are there better more task specific scripting language I can use?

https://redd.it/szqhs6
@r_devops

What is the best way to manage MySQL users

We have an RDS MySQL database and currently users are being created and deleted manually. We are starting to grow and I would like to automate this process.

&#x200B;

I already have an ansible playbook which adds all the developers SSH keys to our VMs, so ideally I would just build off of that and manage MySQL users with the Ansible mysql plugin

&#x200B;

Even with that plugin I am still unsure how to create and distribute the passwords once a user is created?

https://redd.it/szojut
@r_devops