navigator.sendBeacon is a method for POSTing data (string, blob, formData...) to an endpoint asynchronously
The browser guarantees to run these requests to completion, which can be a more stable replacement for onbeforeunload or pagevisibility events with a fetch
Have you used this before?
architecture diagrams before and after migrating the data ingestion pipeline of baselime from aws to cloudflare
I wrote a blog post about how why we migrated, how we did it, how we improved performace, and how much we saved
try to guess how much, you're not ready
A good architecture allows major decisions to be deferred (to a time when you have more information). A good architecture maximizes the number of decisions that are not made. A good architecture makes the choice of tools (database, frameworks, etc.) irrelevant. Plug in the simplest thing that works. If the framework is at the center, you have no architecture at all. You have tools lying in a heap on your workbench with no concept of the thing those tools will build.
(Inspired by @unclebobmartin's talk at https://t.co/gJxSBjUXqt).
Forum PHP, J-11: encore un tout petit peu de patience pour retrouver nos amis de @theodo lors du grand rassemblement de la communauté PHP. Venez à leur rencontre sur leur stand pendant l'événement !
🩶 Theodo, sponsor Argent du Forum PHP 2024 - 10&11 octobre, Disneyland Paris
Cloudflare has officially joined OpenNext to help solve the nextjs vendor lock-in problem
they are now maintaining an official adapter that can deploy your nextjs app to cloudflare
we have a new discord, new github org, and we're adding support in sst soon
Pending their legal claims and litigation against https://t.co/B0aO4F9PVs, WP Engine no longer has free access to https://t.co/B0aO4F9PVs’s resources. https://t.co/SNGtKRJe6w
#WordPress
Pest is not only very elegant on how we write tests, but also helps us on keeping code with the Architectural decisions we prefer.
Check this example 👇
Amazon AWS S3 is probably the largest filesystem in the world, storing 250-1000 Exabytes of data and 100M qps.
Andy Warfield wrote about building this by distributing "heat" with Reed-Solomon erasure coding, Rust-based formal verification and more.
Must-read for engineers.
Yes ! :) Bravo la DreamTeam !
Ce truc est une tuerie :)
On a réécrit 25 ans d’OVHcloud en version « mini » (qui tient dans qq containers K8S avec 16c/24Go) qui permet de gérer avec une API entre 1 et 100+ baies par client, 8 à 5000+ serveurs, la fabric network Spine/Leaf 400G/100G ( 4M de sous réseaux ), et avec l’IAM, le chiffrement native de disques SED, Secure Boot, le stockage de clefs dans le KMS, l’observabilité, le backup et avec la gestion de HW, IPMI, l’état de l’infra, le stock de serveurs, les maintenances, les mises à jour. Chaque client a son Control Plane avec tous les composants.
Il nous reste à finaliser la certification de cette nouvelle plateforme « Bare Metal Pod » avec SecNumCloud mais en parallèle, on attaque le Public Cloud 3AZ avec SNC qui va tourner dessus :)
En plus, dans 3 mois, on commence deployer ce produit « at Edge » et dans les DC de nos clients « On-Prem » avec minimum 1 baie et max 100 baies par projet. Le Bare Metal arrive « clef en main » (HW+SW cablé) avec le nouveau CloudStore qui permet de deployer et gérer VMware, Nutanix, OpenStack, Ceph, OpenIO, MinIO, MySQL, PGSQL, MontoDB, Redis, etc. 40 produits Public Cloud !
En gros, on parle d’une plateforme Public Cloud, 40 produits, en version connecté ou déconnecté, qu’on déploie où on veut:
- une agence bancaire
- un supermarché
- un entrepôt logistique
- une plateforme pertroliere
- un champ d’operation militaire
- un datacentre secret
- un satellite
:)
meet James Hamilton
he’s Amazon’s literal very top engineer, the brain behind all the data centers
he lives in a custom yacht and does not give a fuck about any RTO
NEW POST
GenAI can write code, but is better for understanding legacy code. My colleagues Alessio Ferri, Tom Coggrave, and Shodhan Sheth write about their experiences doing this, including a tool combining AST-fueled knowledge graph and LLMs.
https://t.co/AdQdhk29or
📰 "The Hidden cost of frontend framework" by Oliver Hughes
👍 Nice article reminding us a fwk/tool decision should also be done in the context of maintenance over time
https://t.co/if5eXQvfR5