MS Excel now is Turing complete

It means that it's time to return MS Excel to your resume and write next pet project on really popular language :)

#programming

Разбор падения Slack от 4 января:

https://slack.engineering/slacks-outage-on-january-4th-2021/

Весьма полезное чтиво – хронология, детали, выводы. Кроме ставшего классическим /proc/sys/fs/file-max, есть и специфичные амазоновские причины.

Масштабирование AWS Transit GateWay (TGW)

TGW менеджится Амазоном, потому повлиять на него мы не можем. В то время, как часть проблем у Slack возникла из-за того, что резко возросший трафик через их корневой TGW, через который завязаны их окружения, давал ошибки, не успевая масштабироваться, добавляя проблем во время падения Slack. Амазоновцы вручную боролись с этой ситуацией:

However, our TGWs did not scale fast enough. During the incident, AWS engineers were alerted to our packet drops by their own internal monitoring, and increased our TGW capacity manually.

Чтобы такого избежать, нужно "прогревать" TGW, аналогично тому, как такое предусмотрено для ELB:

https://aws.amazon.com/articles/best-practices-in-evaluating-elastic-load-balancing/#pre-warming

Shared VPC vs different VPCs

Другой момент – отрицательные стороны от использования отдельных VPC. Если бы у Slack использовалась Shared VPC – и для окружения, и для мониторинга, то трафик бы не упёрся бы в узкое горлышко TGW (его скорости масштабирования), через который и соединяются отдельные VPC.

#TGW #Shared_VPC #design

​​A bit of original content for you today.

I’ve wrote an article about my brief experiments with Crossplane.

This is a toolset, which allows you to manage infrastructure as Kubernetes objects.

I promised to write it up last year, but got a chance to actually publish it just now.

Hope you enjoy the reading!

#kubernetes #iac

HCP Consul, a fully managed service mesh, is now generally available on AWS.

It means that you can create a Consul cluster with HashiCorp Cloud Platform, which will run in your AWS cloud environment.

Prices are currently available for Development clusters and start at $0.027/hour. New users will get a free month offering.

#hashicorp #aws

Interesting thread about Nginx Service Mesh by Daniele Polencic

So yeah, Nginx had a service mesh too!

- Read on Threadapp
- Read on Twitter

I'm not sure, if anybody gonna use it for real, though

#kubernetes #nginx #networking

​​A list of awesome Prometheus alerts, divided by the nature. I.e. there are alerts for data bases, proxies and load balancers, storage, etc.

You can just copy-paste these into your monitoring code. Just keep in mind that the thresholds may be different for your particular case!

#observability #monitoring #prometheus

Если кто не знал, то завтра стартует конференция Fosdem`21.
Держите ссылку, смотрите расписание
http://amp.gs/5PTs

​​iamlive - a tool, which generates a basic IAM policy from AWS client-side monitoring (CSM)

iamlive can generate policies based on aws cli commands that you execute. Also, you can configure it to only append required policies for the failed calls. There is a quick GIF demo in the README

Should be useful for everybody who works with AWS

#aws #iam

And we continue with the tools list.

hcltomd is a tool, which as you may guess, converts Terraform variables from HCL into a Markdown table, which you can put into your documentation or a README file.

It would be useful if you're creating Terraform modules for external customers as well asif the people ouside your team are using your modules.

#terraform #hcl #hashicorp #docs

So, the proposal on generics in Go was accepted

#go #programming

Kind - or Kubernetes in Docker - is a tool that allows you to easily spin up your local k8s cluster and was originally developed to test Kubernetes itself.

However, you can use it for the local development as well. This article describes some caveats regarding using Kind for such kind of things.

#kubernetes

Blog posts about why people chose Nomad over Kubernetes are not that common these days. It makes them even more interesting in my opinion.

Here Karan Sharma describes, how he migrated from Kubernetes to Nomad for his home workloads. He also mentioned that they are planning the same migration at his workplace as well.

I would add just that you have to always strive for a "fit for purpose". And there are definitely a lot of purposes Nomad fits well.

#kubernetes #nomad #hashicorp

Right now HashiCorp is livestreaming their annual HashiTalks Conference

It's gonna be 48 hours of tech talks about HashiCorp projects and it's totally free, no registration is required whatsoever

P.S. It's been almost three years and they still haven't added Kyiv to the HUG map 😡

#hashicorp #event

​​Наш наступний гість — Юра Рочняк.

В студентські часи був активним учасником локальних університетських медіа. Написав одну з найвідоміших статей в журналі КПІшник. Працював у Adobe і Preply. Веде один з найбільших ТГ-каналів про DevOps (@catops), створив український осередок Hashicorp User Group і допомагає в організації DevOps Days Kyiv.

Поговоримо про те, чому мати власне девопс-ком'юніті корисно на співбесідах. Як організувати локальну юзер-групу великої корпорації і для чого це потрібно.

Коли: п'ятниця, 19 лютого, 19:00
Де: голосовий чат @dislike_chat

If you were thinking of getting M1 Macbook for DevOps-ish work, but not sure if everything you need would work there, this Reddit thread

Also, you can use Teddit to get old-school ad-free Reddit experience :)

прямо зараз я онлайн 🙂

Через 30 хв починаємо ефір з Юрою Рочняком. Підключитись і задати питання можна тут: @dislike_chat

From our subscribers:

A rant about Helm

tl;dr:
Helm violates some principles of the Kubernetes architecture:
- it is partially imperative
- it not necessarily provides a good abstraction, since you frequently need to know what's going on in a chart
- Helm becomes a monolith on its own, while each Kubernetes component aims to do a suingle thing, but do it good.

Also, author promotes Kustomize and controller/operator approach for your Kubernetes application lifestyle management

#kubernetes #helm

HashiCorp announced a general availability of the Terraform Cloud Operator, which allows you abstract the infrastructure as Kubernetes resources.

In order to use it you have to have Terraform Cloud subnoscription, so this is might be not that important news per se. However, this shows that Kubernetes is a platform on its own and people alongside with well-known companies acknowledge that. Also, managing your insfrastructure from within a Kubernetes cluster doesn't seem like a wrong order of operations. Such things like Terraform Cloud Operator only proves that.

BTW, you can write-up your use case for potential Terraform OSS Operator in this issue

#kubernetes #hashicorp

Навігація:
00:00 - Вступне слово
02:04 - Соц. активність в університеті
09:31 - Журнал "КПІшник" і секс-карта КПІ
14:05 - Створення DevOps-ком'юніті
20:50 - Як ростити спільноту без реклами
27:20 - Мовне питання в інтернаціональних спільнотах
32:18 - Токсік і як з ним боротись
36:50 - Як своє ком'юніті впливає на кар'єру
45:26 - Монетизація своєї спільноти
49:35 - Hashicorp User Group
53:54 - Як зробити свою HUG
59:30- Конференція DevOps Days

Дивитись на YouTube: https://youtu.be/YiqNuZLbzlg

Спілкуємось зі спікерами тут: https://news.1rj.ru/str/dislike_chat
Анонси майбутніх подій тут: https://news.1rj.ru/str/dislike_unsubscribe
Слухати на вашій улюбленій подкаст-платформі: https://anchor.fm/dislike-unsubscribe

CatOps: https://news.1rj.ru/str/catops
HUG Kyiv: https://www.meetup.com/Kyiv-HashiCorp-User-Group/
DevOps Days Kyiv: https://devopsdays.com.ua/

Argo Workflows v3.0 is out!

An article provides some history of Argo development as well as new features of the v3.0 release:

- New APIs for Argo Events
- Controller High-Availability
- Key-only artifacts make it easier to perform map-reduce operations
- Moving the repository. Now it's argo-workflows, not just argo, which caused a lot of confusion with ArgoCD
- Go modules support

#kubernetes #argo #cicd