home.social

#consul — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #consul, aggregated by home.social.

fetched live
  1. Service discovery is client work — even when a platform hides the wiring.

    New walkthrough: two Quarkus provider instances register in Consul with health checks. The consumer uses SmallRye Stork to discover them by name and round-robin between them. Stop one instance and watch the client adapt.

    Registration, health, selection — three jobs, three owners.

    the-main-thread.com/p/service-

    #Quarkus #ServiceDiscovery #Consul #SmallRyeStork #Java #Microservices

  2. Service discovery is client work — even when a platform hides the wiring.

    New walkthrough: two Quarkus provider instances register in Consul with health checks. The consumer uses SmallRye Stork to discover them by name and round-robin between them. Stop one instance and watch the client adapt.

    Registration, health, selection — three jobs, three owners.

    the-main-thread.com/p/service-

    #Quarkus #ServiceDiscovery #Consul #SmallRyeStork #Java #Microservices

  3. Обновление базы за время смены мастера Patroni

    Меня зовут Ирек Агмалов, я DBA-SRE в Ви.Tech - IT-дочке ВсеИнструменты.ру. Мы обновляли PostgreSQL в кластере Patroni и хотели переключить приложение на новую версию без смены строки подключения и без долгого простоя. Для роутинга у нас уже использовались consul-dns и Patroni, поэтому вместо замены DSN мы попробовали временно взять переключение трафика на себя через записи в Consul. В статье покажу, как перевели реплику на PostgreSQL 18, сохранили синхронизацию через логическую репликацию и переключили master и replica так, чтобы потом вернуть кластер в нормальный режим работы.

    habr.com/ru/articles/1036420/

    #postgresql #patroni #consul #logical_replication #upgrade

  4. Consul for Service Networking
    Enable service discovery, secure communication & traffic control across microservices.
    #DevOps #Consul #ServiceMesh #CloudNative

  5. Consul for Service Networking

    Consul helps services find and communicate with each other.
    It provides service discovery, health checks, and secure networking for distributed systems.

    #consul #devops #servicemesh #cloud #microservices

  6. 🎉 Ứng dụng Mac Consul: chỉ cần đổi phần mở rộng, nó tự động chuyển đổi file (hình, video, âm thanh, ebook, config…) hoàn toàn offline. Hỗ trợ chuyển đổi lossless, cả chuyển loại (video→audio, GIF→video). Dùng thử 14 ngày, sau đó mua bản perpetual + 1 năm cập nhật. Thử ngay! #Mac #Consul #FileConversion #TiệnÍch #MacOS #CôngCụ

    reddit.com/r/SideProject/comme

  7. Really annoyed that with #Hashicorp/ #IBM's rogue licensing change, it has killed all interest I had in exploring #Nomad as an alternative to #Kubernetes. It had so much promise. Also, #Consul.

  8. Really annoyed that with #Hashicorp/ #IBM's rogue licensing change, it has killed all interest I had in exploring #Nomad as an alternative to #Kubernetes. It had so much promise. Also, #Consul.

  9. Как мы реализовали георезервирование инфраструктуры для системы видеоконференций: опыт, ошибки, выводы

    Что общего у систем видеоконференцсвязи (ВКС), финансовых транзакций и авиаперевозок? Все они должны работать всегда, когда нужны людям. Сегодня расскажем, как мы строили георезервирование для инфраструктуры, что пошло не так и какие выводы сделали. Перед вами — true story, как мы помогаем одному из наших заказчиков сопровождать инфраструктуру большого критичного сервиса. Привет, Хабр! Это Никита Турцаков и Алексей Кузьмин из К2. В этой статье расскажем: • зачем и для чего нужно георезервирование для системы ВКС; • как мы подошли к выбору архитектуры; • с чем столкнулись при реализации; • и что бы сделали по-другому, если бы начинали сейчас. Готовьтесь: будет много практики и немного боли — потому что как без неё в масштабных и амбициозных задачах.

    habr.com/ru/companies/k2tech/a

    #облачные_сервисы #инфраструктура #облачные_хранилища #цод #георезервирование #кластер #базы_данных #архитектура #consul #postgresql

  10. Загадка от Жака Фреско: как построить свой Rate Limiter и не утонуть в море компромиссов

    Построить Rate Limiter — легко. Сделать его быстрым, отказоустойчивым и работающим в нескольких дата-центрах — сложнее. Делюсь опытом реализации нашего облачного Rate Limiter в DDoS-Guard: принцип работы, анализ правил и реальные примеры из практики.

    habr.com/ru/companies/ddosguar

    #ratelimit #rate_limiting #ratelimiter #rate_limiter #ddosguard #оптимизация_запросов #429_request #облачные_решения #redis #consul

  11. The #s390x open source software team at IBM confirms the latest versions of various software packages run well on #Linux on #IBMZ & #LinuxONE 🐧

    In April 2025 validation was maintained for over 40 projects, including #Apache Hbase, HashiCorp #Consul and #nginx 🎉

    Plus, community CI was added for zaars, V (programming lang), & Exiv2

    Full report + how your project can apply for a s390x VM: community.ibm.com/community/us

  12. The #s390x open source software team at IBM confirms the latest versions of various software packages run well on #Linux on #IBMZ & #LinuxONE 🐧

    In April 2025 validation was maintained for over 40 projects, including #Apache Hbase, HashiCorp #Consul and #nginx 🎉

    Plus, community CI was added for zaars, V (programming lang), & Exiv2

    Full report + how your project can apply for a s390x VM: community.ibm.com/community/us

  13. now does compaction. Using s3 condition writes. 😎

    is good, but complex when not in infra. Especially for . Too many moving parts. 😅

    Very happy litestream is getting some love. Makes even more of a no brainer choice. 😏
    fly.io/blog/litestream-revampe

  14. #Litestream now does compaction. Using s3 condition writes. 😎

    #consul is good, but complex when not in infra. Especially for #litefs. Too many moving parts. 😅

    Very happy litestream is getting some love. Makes #sqlite even more of a no brainer choice. 😏
    fly.io/blog/litestream-revampe

  15. 🚀 I migrated my Mastodon instance from DigitalOcean to my self-hosted Nomad cluster!

    ✅ Lower costs
    ✅ Much faster performance
    ✅ Full control

    I wrote about my setup, including Nomad, Consul, Caddy, and Cloudflare Tunnels.
    🔗 Read the full post: patreon.com/posts/self-hosted-

    Subscribers get ready-to-deploy config files!

    #SelfHosting #Mastodon #Nomad #Consul #Cloudflare #Caddy #DevOps #Infra #Homelab #HashiCorp #FOSS #OpenSource #Linux #SelfHost

  16. 🚀 I migrated my Mastodon instance from DigitalOcean to my self-hosted Nomad cluster!

    ✅ Lower costs
    ✅ Much faster performance
    ✅ Full control

    I wrote about my setup, including Nomad, Consul, Caddy, and Cloudflare Tunnels.
    🔗 Read the full post: patreon.com/posts/self-hosted-

    Subscribers get ready-to-deploy config files!

    #SelfHosting #Mastodon #Nomad #Consul #Cloudflare #Caddy #DevOps #Infra #Homelab #HashiCorp #FOSS #OpenSource #Linux #SelfHost

  17. Vor diesem Hintergrund finde ich immer auch "#Bürgerbeteiligungsportale" leicht absurd, egal ob #Mängelmelder oder #Consul (github.com/consuldemocracy/con)

    Absurd daran finde ich u.a. so Sachen wie, a) dass das typischerweise immer projektspezifisch oder bezogen auf eine spezifische Beziehung zwischen Exekutive und Bürger*innen zum Einsatz kommt, b) politische Vereine und Parteien bestenfalls als "Gäste" am Tisch sitzen usw. Mir schwebt eher vor, sowas zum demokratischen Alltagstool zu machen 6/x

  18. Vor diesem Hintergrund finde ich immer auch "#Bürgerbeteiligungsportale" leicht absurd, egal ob #Mängelmelder oder #Consul (github.com/consuldemocracy/con)

    Absurd daran finde ich u.a. so Sachen wie, a) dass das typischerweise immer projektspezifisch oder bezogen auf eine spezifische Beziehung zwischen Exekutive und Bürger*innen zum Einsatz kommt, b) politische Vereine und Parteien bestenfalls als "Gäste" am Tisch sitzen usw. Mir schwebt eher vor, sowas zum demokratischen Alltagstool zu machen 6/x

  19. #Consul templates have a bunch of functions, but what they don't have is an urlencode function. It has a useless base64urlencode, but then better hope you have control over the code to do a nonstandard urldecode!

    An hour of my life wasted yet again where in the end I just mangle password generation to strip out "bad" characters because the DSN can't handle them and I can't urlencode.

  20. Je découvre Tricot de l'association #Deuxfleurs , un reverseproxy qui utilise #consul et ça m'a l'air bien génial. Ils font aussi le serveur S3 #Garage (que j'utilise fredix.xyz/2024/06/garage/ )

    Les 2 outils sont en #Rust mais ils font aussi du #Go (voir bottin un serveur LDAP)

    guide.deuxfleurs.fr/infrastruc
    guide.deuxfleurs.fr/infrastruc
    guide.deuxfleurs.fr/infrastruc

  21. I've upgraded one of our PostgreSQL test clusters thinking the out-of-cycle release was already there, which wasn't, and broke Patroni because the Python Consul library wasn't updated for years on our systems. I know what I will do next week 😅

    #postgresql #patroni #consul #python

  22. I've upgraded one of our PostgreSQL test clusters thinking the out-of-cycle release was already there, which wasn't, and broke Patroni because the Python Consul library wasn't updated for years on our systems. I know what I will do next week 😅

    #postgresql #patroni #consul #python

  23. Need some advice.. I have my #Nomad cluster working fine, but I don't have any proper load balancing or ingress set up. It's all running on some Debian-based 1L PCs.

    I want a solution to route traffic to them that ticks the following boxes:

    - Local access while on my LAN
    - Remote access (#tailscale)
    - Ideally using hostnames on the LAN
    - Some services connected in LAN without Tailscale (ie. from other devices not capable of connecting via tailscale, or from the wife's machine)

    I suppose I could run a separate box that does #nginx load balancing, as well as the tailscale per-service, but then I'd have a mix of hostnames I can configure each service on.

    Could someone help point me in the right direction? Do I just forgo the tailscale setup and do everything with local hostnames that can be resolved on my tailnet?

    I use a #pihole at home as my first DNS point of contact - could I wire in the #consul DNS to this so I can resolve my services, or is there some better way to do this so I'm passing the nginx-based endpoints to whatever DNS server I'm using? I'm a bit out of my depth on that one.

    #homelab #selfhosting #selfhosted #dns

  24. Need some advice.. I have my #Nomad cluster working fine, but I don't have any proper load balancing or ingress set up. It's all running on some Debian-based 1L PCs.

    I want a solution to route traffic to them that ticks the following boxes:

    - Local access while on my LAN
    - Remote access (#tailscale)
    - Ideally using hostnames on the LAN
    - Some services connected in LAN without Tailscale (ie. from other devices not capable of connecting via tailscale, or from the wife's machine)

    I suppose I could run a separate box that does #nginx load balancing, as well as the tailscale per-service, but then I'd have a mix of hostnames I can configure each service on.

    Could someone help point me in the right direction? Do I just forgo the tailscale setup and do everything with local hostnames that can be resolved on my tailnet?

    I use a #pihole at home as my first DNS point of contact - could I wire in the #consul DNS to this so I can resolve my services, or is there some better way to do this so I'm passing the nginx-based endpoints to whatever DNS server I'm using? I'm a bit out of my depth on that one.

    #homelab #selfhosting #selfhosted #dns

  25. More progress on the #homelab renovation.. mostly software-related at the moment though. Managed to get #nomad to correctly mount the NFS volumes I needed for #smokeping and #mariadb, and figured out how template files work for simple config files being passed in to containers.

    I'm currently using my old #unraid server as the storage backend.. but I'll most likely still be upgrading to a new unraid server soon so what I setup now can easily be translated to the next server. Once I have everything transferred I'll setup a backup task across the storage on unraid so that I have all application states protected.

    Last thing to figure out is the load balancing for my services. I want them all accessible via IP/port on my LAN and by service under #tailscale. Not sure how this will all fit together - I probably need a dedicated box for the LB.. maybe just a simple #nginx proxy with #consul configuration so that it's somewhat dynamic. Would love to get feedback on this, and if you know of a simpler/better way!

  26. More progress on the #homelab renovation.. mostly software-related at the moment though. Managed to get #nomad to correctly mount the NFS volumes I needed for #smokeping and #mariadb, and figured out how template files work for simple config files being passed in to containers.

    I'm currently using my old #unraid server as the storage backend.. but I'll most likely still be upgrading to a new unraid server soon so what I setup now can easily be translated to the next server. Once I have everything transferred I'll setup a backup task across the storage on unraid so that I have all application states protected.

    Last thing to figure out is the load balancing for my services. I want them all accessible via IP/port on my LAN and by service under #tailscale. Not sure how this will all fit together - I probably need a dedicated box for the LB.. maybe just a simple #nginx proxy with #consul configuration so that it's somewhat dynamic. Would love to get feedback on this, and if you know of a simpler/better way!

  27. I’ve cooled off. Got the bugs with #consul and #nomad solved at least.. perhaps I’ll get the CSI plugins for NFS working today? Who knows 😅

  28. Dammit I really don't want to fight this anymore. Thought after the last failed attempt of installing and running a #nomad and #consul cluster that this'd be easier.. this time round. It isn't. It's just failure after crash after conflicting or incomplete docs.

    I don't want to go back to #k8s or #k3s but that line is fast approaching.

  29. Anyone familiar with #hashicorpnomad or #hashicorpconsul know how to fix my broken healthcheck for a client node?

    It says it's not enabled when it is, and even #nomad itself sees that it is. It's only #consul that's failing.