What is our goal?
Performance is and has always been something cursed in some way. Everyone knows that it should be important but nobody tries to do it. Well, we would like to change this!
Performance hub is the source of all tooling that we are working on. We hope that these tools will give you enough confidence to start performance tuning your software. Well as we are mostly practical folks lets move to what is currently included in performance hub.
Database Manipulation Tool
Imagine the situation when you are working on a product that supports multiple database engines and if you would like to perform some tuning. Well, that sucks! You have to create different loads because most of the databases use different dialects. This is certainly not something that will encourage you in this action. Believe us we have been in this situation. Gladly we have a solution for you!
DMT (database-manipulation-tool)
is a Quarkus application that allows you to specify data load in JSON and the tool will take care of the rest. So far we have implemented PostgreSQL, MongoDB, and MySQL dialects and we aim for more! Don’t worry we didn’t forget about the perf part. This tool is currently developed in two modes performance and production. The performance one is stripped of all things that could influence The results of the performance tests or tuning. Production one is more user-friendly, specifically the purpose of integration tests.
Load generator
Well, distributing load into databases is one thing but where we will get the load? Fortunately, this topic is also something that we thought about. We have come up with Load Generator. This generator will effectively create load in the selected pattern. Sometimes you would like constant load, like 500 requests per second for 1 hour. Other times you need to create peak load, like when you want to simulate traffic on the website when tickets to a new concert are released, and many more. Currently, we have only some of these but we will extend the functionality in no time! Database manipulation tool and Load generator are using same JSON schema which means they can and should be used together!
Automation
Besides the previous two applications, we are also storing all the automation we prepare during our performance tuning and testing. Currently, you might find the following helpful ansible scripts there:
- DMT deployment
- MySQL database deployment and auto-tune
- NetData parents and children nodes deployment
In addition to that we have already process automation scripts for PostgreSQL and MongoDB databases.
Monitoring
Lastly and maybe most importantly we are trying to find the most effective solution when it comes to monitoring applications during performance testing or tuning. You might think that this is easy, deploy Prometheus, and problems are solved. This time we will prove you wrong! Most of the monitoring solutions are scraping all the metrics each second (minimal value) and this is certainly not fast enough! Most of the problems happen below one second interval and you will lose all that information because of the scraping.
This problem is not easy to solve and it is surely something we would like to solve in the future.
Join us!
All our projects are open-source under MIT licenses and we are more than happy to welcome new contributors from all around the globe! Doesn’t matter where are you working or how skilled you are simply open a pull request, issue, or write a blog post, we appreciate everything.
