My microservicing concept

Hi guys,I started developing my own microservicing framework since I did not find anything that quite fit my needs.What I wanted was realtime communication between my services and good monitoring and fail detection with retry.First thing I really did not like about existing solutions was HTTP.I think establishing connection every time is overhead for no reason.Instead, I have decided to use redis cluster where I have persistant connection to it and socket is ready to send data at any moment.I have decided to make it more-less basic and only features I created by default are timeout, fail detection and retry (with TTL).For monitoring, I made sure anything worker does is sent to messaging channel so, my separate monitoring service can track state of every single job.In case of failure of my framework, monitor would recreate that job to make sure task is completed (finished or failed).To really use as much of resources as possible, my framework has ability to start multiple node processes and forward them tasks.I was wondering if anyone here had any other ideas for what I could do?Also, job of monitor would be to log any errors with full stack of service calls and arguments.http://ift.tt/2sLJFGV info: http://ift.tt/2EXXYwP

Submitted February 23, 2018 at 01:32PM by nemanjan00