New: Adaptive Concurrency
The queue-worker probes to find the best level of concurrency for your functions for maximum throughput.
#serverless#faas#etl#eventdriven
https://t.co/4LbaHPw1Cu
Introducing our new Python SDK for driving the OpenFaaS REST API including IAM and the Function Builder
#serverless#portable#faas
https://t.co/uN010YY7bn
We added @slicervm support to the K8s Cluster Autoscaler to launch worker nodes in microVMs
Harden your operator locally or in CI:
✅ Scale massively across many machines
✅ Simulate spot instance termination/draining
✅ Run e2e tests in GitHub Actions with nested virt
Where do functions fit in with Claw and Claude? Stable, energy efficient, deterministic automation.
This month we shipped Adaptive Concurrency.
The system figures out optimal load on its own. Big deal for long running tasks and ETL workloads.
https://t.co/sHc6hxfNp0
We've launched a new design for SlicerVM - a fun way to learn and experiment with Firecracker over the holidays. With any commodity hardware or bare-metal cloud like Hetzner.
If you're still wondering about Slicer - here's a quick walkthrough of me booting up a VM with Firecracker and installing @openfaas Edge
https://t.co/jGjItcCiul
Live demo of our new code review bot (opencode + slicer) for GitHub - it's faster, and more flexible than the built-in option.
Would you be interested in testing it out?
https://t.co/TX5atOVyRd
New set of images available for https://t.co/XFQtU6sp2B published on the docs, along with new examples using isolated AI coding agents, VSCode, and a remote Docker builder over SSH.
We took a look at common wisdom and customer feedback on how to optimise AWS EKS/EC2 costs with OpenFaaS.
#faas#costsavings#lambda
https://t.co/yoUGlWXMFY
Queue-Based Scaling is ideal for long-running jobs and background processing - we planned it out in Q4 2024 - and now it's here! 😀 https://t.co/uaIZDwSaWZ
1/3 I patched the Cluster Autoscaler to work with Slicer (manages Firecracker VMs) and K3s.
My functions node pool for @openfaas scales to zero and up whenever needed.
(Hardware below)
The Headroom Controller adds low-priority Pods as a buffer when using autoscaling nodes in Kubernetes
Don't wait 1-2 mins for a new node - pay a tiny bit more for an instant start for your Pods.
https://t.co/mQ1crk1hP1
Build vs. buy - extensibility with customer functions - the dev longs to create their own Heroku, the manager wants to get to market as fast as possible.
Of course we're biased and will do whatever we can so you can ship the capability ASAP.
https://t.co/m0v8KVZETO
We've been seeing more and more cold emails getting through Google's spam filter.. so we took matters into our own hands with @openfaas and @OpenAI
Link in next message..
New example UI for the @openfaas Function Builder API:
✅ Accept JavaScript from customers
✅ Invoke it from your existing product instead of using external webhooks
✅ Code is sandboxed with K8s best practices
#serverless#faas#platform#integration
https://t.co/Jy7y7NnxeA