In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-04-06 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Servers >
Share
Shulou(Shulou.com)05/31 Report--
This article will explain in detail how Knative Serving automatically expands and reduces Autoscaler. The content of the article is of high quality, so the editor shares it for you as a reference. I hope you will have some understanding of the relevant knowledge after reading this article.
By default, Knative Serving provides out-of-the-box fast, request-based automatic capacity expansion-Knative Pod Autoscaler (KPA). Here's how to play Autoscaler in Knative.
Autoscaler mechanism
Knative Serving injects a QUEUE proxy container (queue-proxy) for each POD, which is responsible for reporting user container concurrency metrics to Autoscaler. After receiving these metrics, Autoscaler will adjust the number of POD of Deployment according to the number of concurrent requests and the corresponding algorithm, so as to achieve automatic expansion and reduction.
Arithmetic
Autoscaler expands the capacity based on the average number of requests (concurrency) per POD. The default number of concurrency is 100. Number of POD = total number of concurrent requests / concurrent number of containers
If the number of concurrency in the service is set to 10, Autoscaler will create 5 POD (50 concurrent requests / 10=POD) if the service with 50 concurrent requests is loaded.
Autoscaler implements scaling algorithms in two modes of operation: Stable (stable mode) and Panic (panic mode).
Stable mode
In steady mode, Autoscaler resizes the Deployment to achieve the average number of concurrency required for each POD. The number of concurrency of POD is calculated based on the average number of all data requests received within the 60-second window.
Panic mode
Autoscaler calculates the average number of concurrency within a 60-second window, and the system needs 1 minute to stabilize at the desired level of concurrency. However, Autoscaler also calculates a 6-second panic window, and if the window reaches twice the target concurrency, it will enter panic mode. In panic mode, Autoscaler works on shorter, more sensitive emergency windows. Once the emergency lasts for 60 seconds, Autoscaler returns to the initial 60-second stability window.
| | Panic Target--- > +-- | 20 | +-|-| 10 CONCURRENCY | |
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 0
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
© 2024 shulou.com SLNews company. All rights reserved.