We continuously strive to improve everything we deliver, including our own website. Learn from this case study on how we are optimizing our web fonts to improve the overall web performance.
https://t.co/ineGOUkuQS
#frontend#websiteperformance
@wimsymons @NetcentricHQ We tend to receive very good results for such user numbers running with 28Gb for heap on an m4.4xlarge AWS instance with segment store sizes still fitting in "left" over system memory.
@wimsymons @NetcentricHQ In large setups I have not seen these long GC cycles heavily impacting the performance.G1GC tends to work well or CMS itself also.For 50 conc users a sizing with 64Gb should be sufficient but in the end it does come down to your segment store and that this can be mapped well.
@wimsymons @NetcentricHQ@unicornless already provided good input in general - try to get your segment store memory mapped. This means you need to check how much space your segment store tends to consume after a couple of weeks usage and have this memory free besides your heap.