Hi Nikhil,
Thanks to Jason for the explanation of the first two questions.
I made some replies to the third question, and I also added the slides we discussed before in the attachment.
In the beginning, Jifa only included a worker and frontend two modules, worker provides analysis services,
frontend obtains data through HTTP protocol, and then displays the results. Technically, worker mainly uses
vert.x framework, frontend uses VUE and some are easier Handy UI framework (we found that there will be
bugs in some users' browsers:( ).
Later, in order to be able to serve more users at the same time, we introduced the mater module.
The purpose of mater is to manage workers and dispatch requests to the corresponding workers.
Recently, we created a new branch'master_k8s' on GitHub, hoping that master will use k8s standard API to
dynamically manage workers.
I am very glad that you are willing to try Jifa, although Jifa still has a lot to improve. Any suggestions and
contributions are welcome.
Hope it helps you.
Best,
Denghui
------------------------------------------------------------------
From:Nikhil Dongre <nikhil.vilas@xxxxxxxxx>
Send Time:2021年9月9日(星期四) 13:56
To:jifa developer discussions <jifa-dev@xxxxxxxxxxx>
Subject:[jifa-dev] Query
Hi Team,
We are trying to automate JIFA ( Upload Hprof, Analyse and Report). Our Upload tool will monitor the folder and once hprof generated it will upload to JIFA workspace along with JSON.We can initiate JIFA API to analyse the hprof. I have few question
1. How does cache work?
2. Once Index files are generated, how much memory (RAM) do we need to see Leak suspect or dominator tree reports?
( same as to generate Index files? ex: 70GB Hprof need 90GB RAM?)
3. Can you please share rough high level Architecture for JIFA?
Regards
Nikhil Dongre
