i tried benchmark scotty test network i/o efficiency , overall throughput.
for set 2 local servers written in haskell. 1 doesn't , acts api.
code same is
{-# language overloadedstrings #-} import web.scotty import network.wai.middleware.requestlogger import control.monad import data.text import control.monad.trans import data.bytestring import network.http.types (status302) import data.time.clock import data.text.lazy.encoding (decodeutf8) import control.concurrent import network.http.conduit import network.connection (tlssettings (..)) import network.http.client import network main = scotty 4001 $ middleware logstdoutdev "/dummy_api" $ text $ "dummy response" i wrote server calls server , returns response.
{-# language overloadedstrings #-} import web.scotty import network.wai.middleware.requestlogger import control.monad import control.monad.trans import qualified data.text.internal.lazy lt import data.bytestring import network.http.types (status302) import data.time.clock import data.text.lazy.encoding (decodeutf8) import control.concurrent import qualified data.bytestring.lazy lb import network.http.conduit import network.connection (tlssettings (..)) import network.http.client import network main = let man = newmanager defaultmanagersettings scotty 3000 $ middleware logstdoutdev "/filters" $ response <- liftio $! (testget man) json $ decodeutf8 (lb.fromchunks response) testget :: io manager -> io [b.bytestring] testget manager = request <- parseurl "http://localhost:4001/dummy_api" man <- manager let req = request { method = "get", responsetimeout = nothing, redirectcount = 0} <- withresponse req man $ brconsume . responsebody return $! with both these servers running, performed wrk benchmarking , got extremely high throughput.
wrk -t30 -c100 -d60s "http://localhost:3000/filters" running 1m test @ http://localhost:3000/filters 30 threads , 100 connections thread stats avg stdev max +/- stdev latency 30.86ms 78.40ms 1.14s 95.63% req/sec 174.05 62.29 1.18k 76.20% 287047 requests in 1.00m, 91.61mb read socket errors: connect 0, read 0, write 0, timeout 118 non-2xx or 3xx responses: 284752 requests/sec: 4776.57 transfer/sec: 1.52mb while higher other web servers phoenix, realized meant nothing majority of responses 500 errors occuring due file descriptor exhaustion.
i check limits pretty low.
ulimit -n 256 i increased these limits to
ulimit -n 10240 i ran wrk again , time enough throughput had been reduced drastically.
wrk -t30 -c100 -d60s "http://localhost:3000/filters" running 1m test @ http://localhost:3000/filters 30 threads , 100 connections thread stats avg stdev max +/- stdev latency 105.69ms 161.72ms 1.24s 96.27% req/sec 19.88 16.62 120.00 58.12% 8207 requests in 1.00m, 1.42mb read socket errors: connect 0, read 0, write 0, timeout 1961 non-2xx or 3xx responses: 1521 requests/sec: 136.60 transfer/sec: 24.24kb although amount of 500 errors had reduced, not eliminated. benchmarked gin , phoenix , way better scotty while not giving 500 responses.
what piece of puzzle missing? suspect there issue i'm failing debug.
i understand http-conduit has lot these errors , http-client library uses under hood , has nothing scotty.
@yuras's analogy correct. on running server again, issues related non 2xx status code gone.
the first line in main block culprit. changed line from
main = let man = newmanager defaultmanagersettings to
main = man <- newmanager defaultmanagersettings and voila, there weren't issues. high memory usage of program stabilized 21mb 1gb earlier.
i don't know reason though. nice have explanation this.
Comments
Post a Comment