For an application hosted on Azure cloud. We are trying to get a tool set which is capable of performance testing client side and server side.
We are considering Neoload for server side testing & testing web application over HTTP/HTML protocol. My question is that can Neoload be used for client side performance testing similar to Google Lighthouse. Since Neoload can emulate multiple vusers, can we collect client side metrics like TTFB,First Contentful Paint, Largest Contentful Paint, Load time, etc. with Neoload with multiple users?
You need a process change. If you ask, fix, and confirm that your software is scalable for a single user then you will have fixed about 80% of your performance issues. This, during functional testing, is when you should be leveraging Lighthouse, GTMetrix, and other items. I would also recommend finding a good RUM (Real User Monitor) solution that will automatically collect response time data passively along with all off the w3c navtiming metrics for your pages/apps.
You will find pages that are too heavy, missing indexes on queries, busted cache plans, etc... all with a single user. Fix them at this point, closer to the point of introduction. Far too many organizations wait until it is functionally complete for one user before asking any questions about how fast something is. This allows for the accumulation of performance technical debt in the app/site. In almost all cases the engineering and design decisions are so difficult and costly to unwind late in the process that the app ships knowing there are performance issues in the front end and site design. The "performance hardening spring" will never arrive. This has to be fixed with the next "dot" release
Once, and only once, your software both works for one and scales for one should you then ask if it scales for two or more. Scalability and response time under load is a question of back end resources, How CPU, Disk, Memory, And Network are being leveraged.