Seeing the state of production in git and parameterising deployments that are very similar

Hi folks,

I was once mentored by a fellow who i am no longer in contact with and he recommended to me that all of the config for production remain in a single repository, as a sort of "source of truth" and I was wondering how that might be done and if it is still (if at all) a best practice?

My current thinking is that all config should live in the repository along with source code and the infrastructure required for that code to run in the cloud, so that all of the code required for a service exists as an atomic unit.

But, this causes problems when we e.g. want to deploy 10 services which are all essentially parametrised versions of each other, that is that the source code is the same but the configuration related to each of the 10 deployments is what functionally separates them. Do we separate this information into 10 repositories? For a concrete example we might have 10 data sources that we want to scrape from and all return the same format for data

I would be grateful for any discussion or resources on the matter.

https://redd.it/es1ipp
@r_devops

Minimal tracking of OS health indicators with alarms

I have a one off project that requires the tracking of:

* memory usage
* load average
* disk utilization

for a linux VM. The minimum functionality is to send notifications (email, Slack message) when those metrics exceed a threshold.

Knee-jerk responses might be to use Nagios, prometheus, SNMP etc. I want something simple with minimum effort (max: 1 day effort). What is your expert recommendation ?

https://redd.it/es07t8
@r_devops

Stopping Python-based backdoor Attacks with zerotrust networking

An overview of how to stop attacks like MechaFlounder with a zerotrust networking approach.

https://vimeo.com/383337193/65c5e52bcd

https://redd.it/es32i6
@r_devops

Automated testing for audio DSP software (VST)?

I have been using automated unit tests with TDD (test driven development) at work for physics software development and it made me wonder if audio DSP software could work with tests/TDD.

Some VST developers I know have posted videos about how debugging can be very time consuming for them, how they have noticed that certain parts of the audio processing are never actually executed (which I think is purely a code coverage issue), etc. which interests me in seeing if there is some way to do automated unit tests for TDD.

VSTs aren't standalone software, they run inside a VST host, such as a DAW (Digital Audio Workstation), almost all of which are graphical programs. So far the only way I have found to test them is to build to .dll file and use the VST in the DAW, which is purely manual testing. In addition some DAWs handle plugins differently.

I was thinking a sort of command line based VST host could suffice for setting parameters on the VST, hosting it, letting audio pass in and out, etc. [I did find Ecasound which apparently has command line support](https://lwn.net/Articles/509958/) but it is Linux only. But that is better than nothing.

I was also thinking I could use a recorder to write the output of the plugin to an audio file if that would somehow be useful in testing.

I was thinking some tests could be done fairly easily, such as:

1. Trigger a button press event, ensure that the state the button refers to is triggered
2. Call the various methods as normally, see if anything crashes/raises errors

However, checking if the audio sounds good seems more difficult because it is hard to quantify purely using tests, and it is subjective.

Does anyone have any ideas here? Thanks

https://redd.it/es3y5t
@r_devops

Looking for feedback on tool to fetch all your SaaS audit logs

Hi all,

I am building a tool that fetches the audit logs from all your SaaS tools (think G Suite, Okta, Github, Zendesk, Dropbox, Salesforce, etc) and pushes them into whatever existing logging pipeline you have (Splunk, ELK, Kinesis, etc). The product is in a beta state and I'm searching for testers who can try it out, provide feedback, and help guide my roadmap. If this sounds interesting to you I would love to chat - you can either direct message me here on reddit or shoot me an email at arkadiy{at}logsnitch.com.

Hopefully this fits the promotion rules in this sub but if not I can delete this post.

Thanks very much,

https://redd.it/es2we6
@r_devops

Do you often SSH into remote systems? What are your use cases? Do you have a plan to make it unnecessary in the near future?



https://redd.it/es5ztp
@r_devops

I'm using a makefile to automate terraform deployments. Does anyone know how to pass a Global Red Hat variable to a makefile so I can pass it via SSH agent forwarding. (code below)

echo "$efs_dns" | ssh -At ec2-user@$$(terraform output bastion-public_ip) 'ssh master.openshift.local "cat > efs_dns.txt"'

So this code block is supposed to get the global linux variable '$efs\_dns', pass it to 'master.openshift.local' jumping through a bastion to get the variable output to a file called 'efs\_dns.txt'. The problem is when it's all said and done efs\_dns.txt has some crap value instead of a dns name.

Edit: It seems like the bad output is a piece of the variable name..the variable is called $efs\_dns well the output text in the file is 'fs\_dns'

https://redd.it/es1chl
@r_devops

Moving from network engineering position to DevOps / software positions

​

Hi,

I'm a 23 years old network engineer with 4.5 years of experience at enterprises, I have the following knowledge:

\- Cisco networking - routers and switches also all the protocols.

\- Python scripting and network automation

\- Linux system administration

(not system admin experience but from network engineering perspective).

\- AWS fundamental knowledge

\- Check Point firewalls

\- Virtualization - VMware

​

These are my certs that I currently hold:

\- CCNP R&S

\- CCNA Security

\- AWS Cloud Practitioner

\- LPI Linux Essentials

I currently want to pursue a career in software development or devops and I'm not sure what is the best path for me.

I think about pursuing CS degree, but I got another feeling of doing AWS Solution Architect associate cert and Comptia Linux+ cert and go more to the "devops route"

(also learrning docker + ansible + jenkins afterwards).

I'm not sure if my best way is to pursue a degree or get certs for a better position (Devops/Software)

Thanks,

https://redd.it/ertlj4
@r_devops

Slow your network down

I wrote a blog post about using `netem` to simulate non-ideal network conditions.


Comments and feedback appreciated :)


Thanks


[slow your network down | bad gateway](https://badgateway.qc.to/slow-your-network-down/)

https://redd.it/errmsc
@r_devops

Process Control Solution

Hey, guys, I'm looking for a central ubuntu 18.04 process control solution with a web GUI that can control (stop, start, restart) services for few servers do you know something like this?

https://redd.it/erskjt
@r_devops

As a software engineering student how do I do DevOps

I am a SE student taking on a Capstone project class where I am assigned to a startup company. I have to figure our how and where to host their single web application and MySQL server. I know each choice is circumstantial so how do I make a decision based on my needs. I don't even know how much storage I need.

https://redd.it/errmjc
@r_devops

what are you devops engineers use Golang for ?

Hey

I have 4 month since I started learning golang because it is so cool to learn and has a lot of good features that other programming languages.

but my questions is what can you use it for as a devops / cloud engineer ? what are the use cases ?

https://redd.it/erri4t
@r_devops

Requesting a roadmap to a Big 4 SRE

Hey Reddit,

​

I've posted here on/off for a couple years now. I'm at a point in my career where I'm starting to specialize in a specific job (monitoring, alerting, and log aggregation) and I'm starting to get pigeonholed into it, but also despite several attempts to negotiate with my employer I have not received a significant base salary increase since I joined. Furthermore there are some unsustainable engineering practices we have that I simply no longer wish to participate in and I think it's best I begin to consider other employment options.

Although I understand the fundamentals of DevOps, I think my current skillset is more suited towards some flavor of SRE. I've identified several new skills and concepts I will need to master in order to be able to efficiently transition to a new role (specifically one at a big company) and I've already began studying the interview process.

I just wanted to ask Reddit in general for some advice: What skills would YOU say could get me a general SRE job at a big company and how competent should I be in them? Also, what skills do I need to master to ACE an interview? Would mastering the DevOps roadmap and data structures/algorithms be enough? And for recruiters out there, are there any skills that would really make a candidate stand out? And what skills do you find candidates are too weak in?

Thank you for taking the time to read and respond :)

https://redd.it/erolk2
@r_devops

Using DevOps for Business Development

I know DevOps is not for this purpose, but what do you guys think about using DevOps for tracking account leads for business development? Like creating Epics for different projects that you are trying to win, and creating child issues linked so everyone can track the conversation progress easily.

Some of our colleagues use excel and powerpoint for tracking BD efforts, but this seems a bit old school.

Hoping for a constructive discussion or ideas to improve on this topic.

https://redd.it/esdzur
@r_devops

Inspec vs Twistlock

I was going through various compliance check tools, inspec was the most popular choice out there. But we do use twistlock in our team for vulnerability scanning and I see a compliance check there too.

Is there much difference between inspec and twistlock compliance checks other than former being opensource where we write compliance checks as controls and latter is a paid one already with all compliance checks present ,we just have to run it against the hosts we want to be tested.

https://redd.it/esflfz
@r_devops

PSA: Accidentally get a big charge in your Azure deve account? Microsoft may waive it if you ask!

I recently left some resources I was testing running in my personal Azure dev subscription.

Racked up about $100 in charges before I remembered and removed the resources.

I contacted Azure support and asked if they could waive the charges since I left them running accidentally. They agreed immediately.

So don't hesitate to contact them if you have some big charges! Probably worth a shot with AWS or GCP but haven't tried personally.

https://redd.it/eseyly
@r_devops

Moving from network engineering position to DevOps / software positions

​

Hi,

I'm a 23 years old network engineer with 4.5 years of experience at enterprises, I have the following knowledge:

\- Cisco networking - routers and switches also all the protocols.

\- Python scripting and network automation

\- Linux system administration

(not system admin experience but from network engineering perspective).

\- AWS fundamental knowledge

\- Check Point firewalls

\- Virtualization - VMware

​

These are my certs that I currently hold:

\- CCNP R&S

\- CCNA Security

\- AWS Cloud Practitioner

\- LPI Linux Essentials

​

I currently want to pursue a career in devops/SRE/software development and I'm not sure what is the best path for me.

I think about starting to pursue a CS degree because I didn't go to college yet, but I got another feeling of doing AWS Solution Architect associate cert and Linux cert and go more to the "devops/SRE route"

(also learrning docker + ansible + jenkins afterwards).

​

I'm not sure if my best way is to pursue a CS degree or get certs for a better position (Devops/Software)

​

Also I'm kind of wondering how to advance my skills in order to become devops engineer / SRE.

If someone can structure for me a learning path in terms of subjects, I can learn and also do labs and test for myself.

Also what are like the best tools for devops? I heard that AWS + Linux + Ansible + Docker + Kubernetes and Jenkins/TravisCI are the most common.

Thanks,

https://redd.it/eshvke
@r_devops

Alternatives to Redgate for SQL Database Automation

we've been using redgate software to automate our sql build and deployment practices for a few years now; however, they changed their license model a couple of years ago and its incredibly expensive now on a per-user basis.

The redgate tooling works very well for us combining sql source control with the DLM automation tools, the flexibility the filter files and comparison options provide is pretty important as we have close to 100 different databases, many of which have some specific deployment requirements to ignore particular objects. The comparison reports it generates are critical as well to provide visibility into changes being made. These are all features that fall into the 'must have' category for us. The other nice thing about redgate is that it actually generates all the code when commits are made so we have consistent SQL code across all our databases.

​

Would love to hear any suggestions you have on alternative tooling or approaches to full automation around the management of SQL Database schema and data.

https://redd.it/eskesa
@r_devops

How to use OperatorSDK and service meshes together?

Here is a quickie one:
How to use concept of an operator and service mesh in Kubernetes based, microservice/event driven infrastructure?
How to make most of both patterns? Are there any patterns/tutorial you can recommend?
Cheers!

https://redd.it/esjp76
@r_devops

Migrating to Helm 3

Hi all,

I've recently finished migrating our environments over to Helm 3 (goodbye and good riddance Tiller!). I've written some blog posts that documented some of the pitfalls I encountered when updating our scripts for Helm 3, as a lot of the flags for Helm have changed, and some don't even exist anymore.

There's also a great blog post by the Helm developers that explains the migration process from a high level. One thing that was missing is what to do if you find yourself wanting to roll back to Helm 2. How do you delete releases once they're migrated to Helm 3? Conveniently enough, release data is stored in k8s secrets now, so it's pretty easy to do.

Anyway, hopefully these are of use to some of you.
[Breaking Changes in Helm 3 (and How to Fix Them)](https://itnext.io/breaking-changes-in-helm-3-and-how-to-fix-them-39fea23e06ff?source=friends_link&sk=c1514964ac1db15574384b234a0c5bdf)
[Migrating to Helm 3: What You Need to Know](https://itnext.io/additional-tips-for-migrating-to-helm-3-304c9d50f1b4?source=friends_link&sk=2aed5a6f6e754369d3887493ed202d8c)

https://redd.it/esj90t
@r_devops

Kubernetes pulling wrong image

I have an issue where kubernetes is pulling the wrong image from ECR. Issue pretty much is serviceA:tagX has Image ID of ABC then serviceB:tagY has an Image ID of ABC as well so Kubernetes is deploying the serviceA image for serviceB. I'm using helm charts and right now am deploying manually (for now) via setting the tag as a values.yaml param. What can I do to prevent this issue from coming up again?

https://redd.it/esmnof
@r_devops