Finished my small contest for major frameworks such @micronautfw , @QuarkusIO , @springframework
in order to compare their @awscloud #AWSLambda implementations for @graalvm native image setups and here we are: github.com/GoodforGod/awslam…
ᴀɴᴛᴏɴ retweeted
Here are some psrecord plots comparing @GraalVM Native Image vs JVM in terms of CPU and memory usage over time, using @pkolaczk's virtual threads benchmark (see quoted tweet).
Memory usage seems to peak at the end, at least on the JVM.
Nonetheless, the winner is quite clear. 🤓
ᴀɴᴛᴏɴ retweeted
📖 Micronaut Framework Code Generation hubs.li/Q01HDm0z0 #micronaut
ᴀɴᴛᴏɴ retweeted
A lot of options to generate a @Docker image of a @micronautfw application. Go as fast 🏎️ as you need to. @AzulSystems CRaC, @graalvm native exec, @java
D. Img: 2s 473ms
Optimized D. Img: 2s 359ms
CRaC D. Img: 462ms
Native D. Img: 47ms
Optimized Native D.Img: 39ms #micronaut
📖 Micronaut Guide - Building a Docker Image of your Micronaut application hubs.li/Q01C5_gy0 #micronaut
The idea of such “toolkit” around Spring only shows how Spring is over complicated, non intuitive and over engineered as a framework.
You properly configured app with framework should do this out of the box with build tool in conjunction, otherwise you already lost.
📢 Introducing JUST 🚀
Just is a command line toolkit for developing @springboot applications:
👉 just.maciejwalkowiak.com/
It's like a superpower for 🍃Spring - makes you not only faster developer but also a happier person!
Lets dive into Just features 🧵
ᴀɴᴛᴏɴ retweeted
Image shows the time to first response for a Docker image of a @micronautfw Application. The app uses Micronaut Data JPA with Hikari connection pool, H2, and the Netty runtime.
Standard / Restored from an @AzulSystems CRaC Checkpoint / Native Executable with @graalvm #micronaut
Would be interesting to measure all main frameworks 🤔
Experimenting with CRaC & @springboot on Jetty (Tomcat is not CRaC friendly)
- startup time without CRaC: 7338ms
- startup time with CRaC: 439ms
Numbers are so bad because it runs on VirtualBox VM, but damn that's almost 20x improvement!
ᴀɴᴛᴏɴ retweeted
Если у человека стоит флажок в нике, то это многократно увеличивает шанс, что перед вами ебанько.
ᴀɴᴛᴏɴ retweeted
Reasons for writing Unit Tests:
1. To verify business logic if you have any
2. To meet 80% code coverage rule
Reasons to write Integration Tests:
1. To verify your code actually works
Use @testcontainers and write real tests.
ᴀɴᴛᴏɴ retweeted
"Why are salaries after tax lower in countries like France/Belgium when the cost of living is comparable to e.g. the UK or Switzerland?"
Because of taxes! And cost of employment for companies. Here's how much it would cost for a company to pay everyone €100K/yr after taxes:
Presenting #GraalVM Hint library that will help you easily and type-safely generate @graalvm native-image configurations files by using annotations.
github.com/GoodforGod/graalv…
This took quite a time for me to design such test, gather data for different setups and tune my logging implementation, you can see the difference vs #slf4j-simpler-logger is 3x and up to 9x in different scenarios!
Definitely can put my hat to #logback developers, they did amazing work on optimizing and tuning logger to performance so greatly and been available to be flexible and configurable at the same time! 🎩
2) Environment logging and logger level refresh. This is nice feature to add some of environmental variables to logging output, quite useful feature in different setups. Also it is useful to change logger level in runtime, which is also not possible in slf4j-simple-logger impl.















