benchmark: v0.4, made a benchmark for `apisix` to determine if there is a performance issue

[apache]

benchmark: v0.4, made a benchmark for `apisix` to determine if there is a performance issue

[membphis]

1 route + 1 upstream + 2 plugin(limit-conn + prometheus)

# route information
$ curl http://127.0.0.1:2379/v2/keys/apisix/routes/1 -X PUT \
    -d value='{"methods":["GET"],"uri":"/hello","id":1,"plugin_config":{"limit-count":{"count":200000000,"time_window":60,"rejected_code":503,"key":"remote_addr"},"prometheus":{}},"upstream":{"type":"roundrobin","nodes":{"127.0.0.1:80":1,"127.0.0.2:80":1}}}'

$ wrk -d 5 --latency  http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   732.47us  159.15us   2.62ms   89.17%
    Req/Sec     6.78k   460.18     7.50k    67.65%
  Latency Distribution
     50%  685.00us
     75%  757.00us
     90%    0.90ms
     99%    1.39ms
  68815 requests in 5.10s, 16.21MB read
Requests/sec:  13492.64
Transfer/sec:      3.18MB

$ wrk -d 5 --latency  http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   770.87us  141.37us   3.86ms   85.88%
    Req/Sec     6.44k   335.48     7.17k    70.59%
  Latency Distribution
     50%  734.00us
     75%  812.00us
     90%    0.92ms
     99%    1.26ms
  65363 requests in 5.10s, 15.39MB read
Requests/sec:  12815.03
Transfer/sec:      3.02MB

flame graph.zip

[membphis]

better version with this commit:
0a94996

$ curl http://127.0.0.1:2379/v2/keys/apisix/routes/1 -X PUT \
    -d value='{"methods":["GET"],"uri":"/hello","id":1,"plugin_config":{"limit-count":{"count":200000000,"time_window":60,"rejected_code":503,"key":"remote_addr"},"prometheus":{}},"upstream":{"type":"roundrobin","nodes":{"127.0.0.1:80":1,"127.0.0.2:80":1}}}'

$ wrk -d 500 --latency  http://127.0.0.1:9080/hello
Running 8m test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
^C  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   680.99us  126.71us   1.91ms   83.71%
    Req/Sec     7.30k   438.79     8.21k    72.73%
  Latency Distribution
     50%  654.00us
     75%  718.00us
     90%  823.00us
     99%    1.20ms
  31941 requests in 2.20s, 7.52MB read
Requests/sec:  14519.97
Transfer/sec:      3.42MB

$ wrk -d 5 --latency  http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   663.87us  148.15us   2.51ms   87.31%
    Req/Sec     7.49k     0.92k   14.09k    87.13%
  Latency Distribution
     50%  634.00us
     75%  697.00us
     90%  823.00us
     99%    1.25ms
  75269 requests in 5.10s, 17.73MB read
Requests/sec:  14758.49
Transfer/sec:      3.48MB

# nothing plugin was enabled
$ curl http://127.0.0.1:2379/v2/keys/apisix/routes/1 -X PUT \
    -d value='{"methods":["GET"],"uri":"/hello","id":1,"plugin_config":{},"upstream":{"type":"roundrobin","nodes":{"127.0.0.1:80":1,"127.0.0.2:80":1}}}'

$ wrk -d 5 --latency  http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   600.46us  163.78us   3.90ms   91.57%
    Req/Sec     8.28k   726.78     9.39k    72.00%
  Latency Distribution
     50%  556.00us
     75%  613.00us
     90%  735.00us
     99%    1.29ms
  82384 requests in 5.00s, 14.37MB read
Requests/sec:  16474.20
Transfer/sec:      2.87MB

$ wrk -d 5 --latency  http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   623.03us  201.04us   4.50ms   87.49%
    Req/Sec     7.99k     1.03k   15.15k    79.21%
  Latency Distribution
     50%  557.00us
     75%  655.00us
     90%  842.00us
     99%    1.40ms
  80322 requests in 5.10s, 14.01MB read
Requests/sec:  15752.11
Transfer/sec:      2.75MB

[membphis]

commit: fe02056 + patch: https://github.com/iresty/lua-var-nginx-module

# 1upstream + 1 route + 2 plugin (limit-count + prometheus)

$ curl http://127.0.0.1:2379/v2/keys/apisix/routes/1 -X PUT \
    -d value='{"methods":["GET"],"uri":"/hello","id":1,"plugin_config":{"limit-count":{"count":200000000,"time_window":60,"rejected_code":503,"key":"remote_addr"},"prometheus":{}},"upstream":{"type":"roundrobin","nodes":{"127.0.0.1:80":1,"127.0.0.2:80":1}}}'

$ wrk -d 5 --latency http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   641.54us  158.44us   3.64ms   93.50%
    Req/Sec     7.78k   515.63     8.71k    58.82%
  Latency Distribution
     50%  603.00us
     75%  672.00us
     90%  756.00us
     99%    1.24ms
  78919 requests in 5.10s, 18.59MB read
Requests/sec:  15475.07
Transfer/sec:      3.64MB

$ wrk -d 5 --latency http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   693.14us  142.28us   3.66ms   89.51%
    Req/Sec     7.20k   396.29     8.10k    75.00%
  Latency Distribution
     50%  660.00us
     75%  731.00us
     90%  816.00us
     99%    1.19ms
  71566 requests in 5.00s, 16.85MB read
Requests/sec:  14312.39
Transfer/sec:      3.37MB

# 1route + 1 upstream + no plugin
$ curl http://127.0.0.1:2379/v2/keys/apisix/routes/1 -X PUT -d value='{"methods":["GET"],"uri":"/hello","id":1,"plugin_config":{},"upstream":{"type":"roundrobin","nodes":{"127.0.0.1:80":1,"127.0.0.2:80":1}}}'

$ wrk -d 5 --latency http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   567.92us  113.97us   3.95ms   92.04%
    Req/Sec     8.69k   476.91     9.31k    68.63%
  Latency Distribution
     50%  542.00us
     75%  584.00us
     90%  656.00us
     99%    0.99ms
  88200 requests in 5.10s, 15.39MB read
Requests/sec:  17293.04
Transfer/sec:      3.02MB

$ wrk -d 5 --latency http://127.0.0.1:9080/hello
Running 5s test @ http://127.0.0.1:9080/hello
  2 threads and 10 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   557.94us   88.09us   2.58ms   87.68%
    Req/Sec     8.82k   420.66     9.71k    76.47%
  Latency Distribution
     50%  533.00us
     75%  583.00us
     90%  650.00us
     99%    0.90ms
  89496 requests in 5.10s, 15.61MB read
Requests/sec:  17548.13
Transfer/sec:      3.06MB

flame graph.zip

[membphis]

Benchmark on google cloud(1 route):

Machine type
You must stop the VM instance to edit its machine type

n1-highcpu-8 (8 vCPUs, 7.2 GB memory)

https://gist.github.com/membphis/9e7e3799c6bc83156faf2f7fe2fade7e

Flame graph:

google_noplugin.svg.zip
google_2plugin.svg.zip