05-profiling/notes.txt

Profiling


Maintainability X Performance

var products []Product

func FilterProducts() []Product{
    var electronicsProducts []Product
    for i:=; i < len(products); i++{
        if products[i].Category == "Electronics" {
            electronicsProducts = append(electronicsProducts, products[i])
        }
    }
    return electronicsProducts
}

electronicsProducts := Products.Filter(func(p){
    return p.Category == "Electronics"
})


func (products Products) Filter(predicate func(Product)bool ) Products {
    var result []Product
    for i:=; i < len(products); i++{
        if predicate(products[i]) {
            result = append(result, products[i])
        }
    }
    return result
}


How?
    Benchmarking

Tools
    go toolset
    
    load testing tools
        ab
        go-wrk


To run benchmark with memory
    go test ./... -bench=. -benchmem

To generate the cpu profile
    go test ./... -bench=. -benchmem -cpuprofile prof.cpu

    commands used in the pprof terminal
        top10
        top10 -cum
        list <function name>
        

To access the profile data using a web interface
    go tool pprof -http=":8080" profiling-app.test prof.cp


To profile a web service
    import _ "net/http/pprof"

load test the service using ab or go-wrk

while the load is still being applied
    visit http://localhost:8080/debug/pprof/


to investigate the trace file
go tool trace sum.trace