Autonomous concurrency in Sedai puts an end to cold starts for Lambdas. Using a combination of provisioned concurrency, Lambda extensions, and traffic seasonality, Sedai aautonomously eliminates cold starts with minimal to no changes in cost.
How it works:
Sedai continuously monitors the traffic patterns for every Lambda and builds a seasonality model to understand local and global trends
Lambda extensions help monitor the behavior of every Lambda, especcially concurreency patterns, runtime events and cold start duration. Written in go as an external extension, this allows monitoring Lambdas up close without any overheads
Sedai's AV lambda, acts as the smart control unit monitoring the Lambdas for concurrency events and scaling concurrency up and down based on Seasonality
Sedai's core provides the central control and learning unit to drive the intelligence in all the components above.
Currently Autonomous concurrency supports the following runtimes:
NodeJS 12 and above on x86_64
Python 3.7 and above on x86_64
Java 8 and above on x86_64
The static and runtime characteristics of the extension are limited to:
Additional 12 MB size in deployed Lambda layer
0.1 ms latency overhead per 1000 invocations
0.11 ms execution time overhead per 1000 invocations