Connection timed out when hitting more than 10000 QPS via gatling,

Gatling (Open-Source)
1

Bandwidth?

2

You mean internet bandwidth to hit the server?

Also, here is my gatling.conf, any suggestions on this :

ahc {

#keepAlive = true # Allow pooling HTTP connections (keep-alive header automatically added)

connectTimeout = 600000 # Timeout when establishing a connection

handshakeTimeout = 600000 # Timeout when performing TLS hashshake

#pooledConnectionIdleTimeout = 60000 # Timeout when a connection stays unused in the pool

readTimeout = 600000 # Timeout when a used connection stays idle

#maxRetry = 2 # Number of times that a request should be tried again

requestTimeout = 600000

#maxRetry = 2 # Number of times that a request should be tried again

requestTimeout = 600000 # Timeout of the requests

#acceptAnyCertificate = true # When set to true, doesn’t validate SSL certificates

#httpClientCodecMaxInitialLineLength = 4096 # Maximum length of the initial line of the response (e.g. “HTTP/1.0 200 OK”)

#httpClientCodecMaxHeaderSize = 8192 # Maximum size, in bytes, of each request’s headers

#httpClientCodecMaxChunkSize = 8192 # Maximum length of the content or each chunk

#webSocketMaxFrameSize = 10240000 # Maximum frame payload size

#sslEnabledProtocols = [TLSv1.2, TLSv1.1, TLSv1] # Array of enabled protocols for HTTPS, if empty use the JDK defaults

#sslEnabledCipherSuites = [] # Array of enabled cipher suites for HTTPS, if empty use the AHC defaults

#sslSessionCacheSize = 0 # SSLSession cache size, set to 0 to use JDK’s default

#sslSessionTimeout = 0 # SSLSession timeout in seconds, set to 0 to use JDK’s default (24h)

#useOpenSsl = false # if OpenSSL should be used instead of JSSE (requires tcnative jar)

#useNativeTransport = false # if native transport should be used instead of Java NIO (requires netty-transport-native-epoll, currently Linux only)

#tcpNoDelay = true

#soReuseAddress = false

#soLinger = -1

#soSndBuf = -1

#soRcvBuf = -1

#allocator = “pooled” # switch to unpooled for unpooled ByteBufAllocator

#maxThreadLocalCharBufferSize = 200000 # Netty’s default is 16k

}

3

I mean bandwidth at any network component between your Gatling injectors and your app servers: servers NIC, network switches, etc.

In any case, this is not a Gatling issue. If you don’t hit a CPU limit, that means you’re hitting an IO limit: either you server doesn’t respond fast enough, or you’re saturating your network.

4

Is there any way in gatling to reuse the connections?