An open cloud native capacity solution which helps you achieve ultimate resource utilization in an intelligent and risk-free way.
Home Page: https://kapacity.netlify.app
License: Apache License 2.0
I followed the tutorial and found that the training was incorrect and promethus did not have the nginx_ingress_controller_requests_rate indicator.
success to train the model
follow the tutorial
No response
$ kubectl version
# paste output hereaccording the wechat article (point this). It says in June. the v0.2 will be publish. but I find it not get. what I more concerned is , in which time. I could get release version in order to use it in prod env.
wait for your reply. thanks.
want to know release version in which time could get
Introduce the Python script to train the time series forecasting model for predictive scaling.
The time series forecasting model need to be trained beforehand so that the acutal prediction job can utilize it.
Add a reconcile-concurrency flag to manager to configure the reconcile concurrency of every controller.
The reconcile concurrency defaults to 1 which does not meet the efficiency requirement in production.
Support Kubernetes CronJob as external horizontal portrait algorithm job.
It is a lightweight way to run complex long-running algorithms.
Support collecting result of external algorithm job from ConfigMap to HorizontalPortrait.
It is a lightweight way to collect result of external algorithm jobs which are able to write ConfigMap in cluster.
Introduce a best-effort pod sort algorithm for Deployment as well as ReplicaSet.
To fully support replica status control.
Currently, the behavior of the portrait selection is undefined when we have multiple portraits with the same priority. We need to introduce a reasonable behavior to it, that is to select the one which desires most replicas.
To provide a more reasonable and predictable behavior.
I followed the tutorial to create a ihpa : https://kapacity.netlify.app/zh-cn/docs/getting-started/quick-start/ihpa/predictive-scaling/#%E5%88%9B%E5%BB%BA%E9%85%8D%E7%BD%AE%E4%BA%86%E5%8A%A8%E6%80%81%E9%A2%84%E6%B5%8B%E5%BC%8F%E7%94%BB%E5%83%8F%E6%BA%90%E7%9A%84-ihpa, and I failed.
success to run the job
No response
$ kubectl version
# paste output hereWhen is the code implementation of the burst algorithm expected to be released?
// TODO
BurstPortraitType PortraitType = "Burst"
“Algorithm which detects abnormal traffic or potential capacity risks, and suggests a safe replica count proactively.”
I originally thought the burst algorithm mentioned in the project introduction was already supported.
Expose the unified metric provider interface as a (probably gRPC) service.
This enables external algorithm jobs to utilize the internal unified metric provider interface so that they don't need to impl metrics query themselves.
The dataset required for algorithm training
This can help our participants conduct algorithmic research and verify the excellence of the Kapacity algorithm
Hello， I learned that kapacity has a dashboard project which will introducted，I want to know when I could get it.
use dashboard as a control plane
Introduce an admission webhook to automatically inject the kapacity.traas.io/online readiness gate to pods which needed.
To support the readiness gate pod traffic controller.
Introduce the Python program which gathers metrics, does replicas prediction, and writes out results. This program can be run as Kubernetes CronJob. The model used by the program should be trained beforehand so that it would not rely on GPUs.
It is needed to do the real prediction work for the predictive portrait.
Support arbitrary metrics query for promethues metric provider.
Currently the Promethues metric provider only supports resource metrics query which limits the metrics that algorithms can use. We need a machinism to support arbitrary PQL query to make most of this provider.
Hello，I want to know, if I can use api（such as http invoke）,to control the kapacity behavior . If it is yes, where could I get the api?
when I could use http method to control kapacity behavior， I could make our own control portal page. and use http api to publish config into k8s cluster.
Thank you very much for open-sourcing Kapacity. Sharing experiences from the industry is very important for this field. My main research direction includes container workload resource recommendation, so even if the resource recommendation algorithm of Kapacity will be open-sourced in the future, I still want to learn about your ideas as soon as possible. Could you please provide a detailed introduction of the models and methods you use for resource recommendation?
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.