- Install
- Benchmarks
- Constant Request Rate Benchmarks
- User Concurrency Benchmarks Without Random Sleep
- User Concurrency Benchmarks With Random Sleep
- bench-ramping-vus.sh automated k6 benchmarking demo 1
- bench-ramping-vus.sh automated k6 benchmarking demo 2 + JSON log to InfluxDB batch line format
- bench-ramping-vus.sh automated k6 benchmarking demo 3 + JSON log to InfluxDB batch line format
- InfluxDB + Grafana
- k6 Log Parsing Into InfluxDB Batch Line Format
- k6 handleSummary JSON Log Parsing
sudo gpg -k
sudo gpg --no-default-keyring --keyring /usr/share/keyrings/k6-archive-keyring.gpg --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys C5AD17C747E3415A3642D57D77C6C491D6AC1D69
echo "deb [signed-by=/usr/share/keyrings/k6-archive-keyring.gpg] https://dl.k6.io/deb stable main" | sudo tee /etc/apt/sources.list.d/k6.list
sudo apt-get update
sudo apt-get install k6
Using dnf (or yum on older versions):
sudo dnf install https://dl.k6.io/rpm/repo.rpm
sudo dnf install k6
k6 version
k6 v0.49.0 (commit/b5328aa782, go1.21.6, linux/amd64
k6 --help
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
Usage:
k6 [command]
Available Commands:
archive Create an archive
cloud Run a test on the cloud
completion Generate the autocompletion script for the specified shell
help Help about any command
inspect Inspect a script or archive
login Authenticate with a service
new Create and initialize a new k6 script
pause Pause a running test
resume Resume a paused test
run Start a test
scale Scale a running test
stats Show test metrics
status Show test status
version Show application version
Flags:
-a, --address string address for the REST API server (default "localhost:6565")
-c, --config string JSON config file (default "/root/.config/loadimpact/k6/config.json")
-h, --help help for k6
--log-format string log output format
--log-output string change the output for k6 logs, possible values are stderr,stdout,none,loki[=host:port],file[=./path.fileformat] (default "stderr")
--no-color disable colored output
--profiling-enabled enable profiling (pprof) endpoints, k6's REST API should be enabled as well
-q, --quiet disable progress updates
-v, --verbose enable verbose logging
--version version for k6
Use "k6 [command] --help" for more information about a command.
Install psrecord
pip install psrecord
psrecord --help
usage: psrecord [-h] [--log LOG] [--plot PLOT] [--duration DURATION]
[--interval INTERVAL] [--include-children]
process_id_or_command
Record CPU and memory usage for a process
positional arguments:
process_id_or_command
the process id or command
optional arguments:
-h, --help show this help message and exit
--log LOG output the statistics to a file
--plot PLOT output the statistics to a plot
--duration DURATION how long to record for (in seconds). If not specified,
the recording is continuous until the job exits.
--interval INTERVAL how long to wait between each sample (in seconds). By
default the process is sampled as often as possible.
--include-children include sub-processes in statistics (results in a
slower maximum sampling rate).
Add a psrecord-to-json.sh tool to allow converting psrecord cpu and memory data into JSON and InfluxDB data formats.
./psrecord-to-json.sh
Usage:
./psrecord-to-json.sh json psrecord_file.log
./psrecord-to-json.sh influx psrecord_file.log
For REQRATE = 100 request/s constant rate test with psrecord. Where --duration $((TIME+30)) is the TIME + 30s defined in benchmark-rps.js.
The spid variable is for Nginx web server's MainPID via Centmin Mod cminfo service-info tool.
cminfo service-info nginx
{
"Names": "nginx.service",
"Description": "Centmin Mod NGINX Server",
"Type": "forking",
"ActiveState": "active",
"LoadState": "loaded",
"SubState": "running",
"Result": "success",
"ExecMainStartTimestamp": "Thu 2022-10-06 12:59:35 UTC",
"MainPID": "28166",
"DropInPaths": "/etc/systemd/system/nginx.service.d/failure-restart.conf /etc/systemd/system/nginx.service.d/mimalloc.conf /etc/systemd/system/nginx.service.d/openfileslimit.conf",
"ExecStart": "{ path=/usr/local/sbin/nginx ; argv[]=/usr/local/sbin/nginx -c /usr/local/nginx/conf/nginx.conf ; ignore_errors=no ; start_time=[Thu 2022-10-06 12:59:35 UTC] ; stop_time=[Thu 2022-10-06 12:59:35 UTC] ; pid=28162 ; code=exited ; status=0 }",
"ExecStartPre": "{ path=/usr/local/sbin/nginx ; argv[]=/usr/local/sbin/nginx -t ; ignore_errors=no ; start_time=[Thu 2022-10-06 12:59:34 UTC] ; stop_time=[Thu 2022-10-06 12:59:35 UTC] ; pid=28158 ; code=exited ; status=0 }",
"ExecReload": "{ path=/bin/sh ; argv[]=/bin/sh -c /bin/kill -s HUP $(/bin/cat /usr/local/nginx/logs/nginx.pid) ; ignore_errors=no ; start_time=[n/a] ; stop_time=[n/a] ; pid=0 ; code=(null) ; status=0/0 }",
"ExecStop": "{ path=/bin/sh ; argv[]=/bin/sh -c /bin/kill -s TERM $(/bin/cat /usr/local/nginx/logs/nginx.pid) ; ignore_errors=no ; start_time=[Thu 2022-10-06 12:59:34 UTC] ; stop_time=[Thu 2022-10-06 12:59:34 UTC] ; pid=28152 ; code=exited ; status=0 }",
"PIDFile": "/usr/local/nginx/logs/nginx.pid",
"LimitMEMLOCK": "65536",
"LimitNOFILE": "1048576",
"LimitNPROC": "127803",
"After": "nss-lookup.target network-online.target remote-fs.target systemd-journald.socket basic.target system.slice syslog.target",
"Before": "multi-user.target shutdown.target",
"Conflicts": "shutdown.target",
"FailureAction": "none",
"FragmentPath": "/usr/lib/systemd/system/nginx.service",
"NotifyAccess": "none",
"PrivateNetwork": "no",
"PrivateTmp": "no",
"ProtectHome": "no",
"ProtectSystem": "no",
"Requires": "basic.target system.slice",
"Restart": "on-failure",
"RestartUSec": "5s",
"StartLimitAction": "none",
"StartLimitBurst": "5",
"StartLimitInterval": "30000000",
"TimeoutStartUSec": "1min 30s",
"TimeoutStopUSec": "1min 30s",
"UnitFilePreset": "disabled",
"UnitFileState": "enabled",
"WantedBy": "multi-user.target",
"Wants": "network-online.target"
}VU=20
REQRATE=100
TIME=30
DOMAIN=https://yourdomain.com/
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME+30)) --log psrecord-rps-nginx.log --plot plot-rps-nginx.png &
psrecord "taskset -c 0-3 k6 run -e RPS=${REQRATE} -e DURATION=${TIME}s -e USERS=${VU} -e URL=$DOMAIN --no-usage-report --out json=summary-raw-rps.gz benchmark-rps.js" --include-children --interval 0.1 --duration $((TIME+30)) --log psrecord.log --plot plot.png
VU=20
REQRATE=100
TIME=30
DOMAIN=https://yourdomain.com/
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME+30)) --log psrecord-rps-nginx.log --plot plot-rps-nginx.png &
psrecord "taskset -c 0-3 k6 run -e RPS=${REQRATE} -e DURATION=${TIME}s -e USERS=${VU} -e URL=$DOMAIN --no-usage-report --out json=summary-raw-rps.gz benchmark-rps.js" --include-children --interval 0.1 --duration $((TIME+30)) --log psrecord.log --plot plot.png
Starting up command 'taskset -c 0-3 k6 run -e RPS=100 -e DURATION=30s -e USERS=20 -e URL=https://domain1.com/ --no-usage-report --out json=summary-raw-rps.gz benchmark-rps.js' and attaching to process
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
execution: local
script: benchmark-rps.js
output: json (summary-raw-rps.gz)
scenarios: (100.00%) 1 scenario, 1000 max VUs, 1m0s max duration (incl. graceful stop):
* open_model: 100.00 iterations/s for 30s (maxVUs: 20-1000, gracefulStop: 30s)
running (0m30.0s), 0000/0020 VUs, 3000 complete and 0 interrupted iterations
open_model ✓ [======================================] 0000/0020 VUs 30s 100.00 iters/s
█ main index page
✓ is status 200
checks.........................: 100.00% ✓ 3000 ✗ 0
data_received..................: 7.1 MB 238 kB/s
data_sent......................: 131 kB 4.4 kB/s
group_duration.................: avg=288.4µs min=232.7µs med=271.13µs max=6.36ms p(95)=330.81µs p(99)=437.09µs p(99.99)=5.79ms count=3000
http_req_blocked...............: avg=5.98µs min=154ns med=259ns max=3.93ms p(95)=360ns p(99)=799ns p(99.99)=3.04ms count=3000
http_req_connecting............: avg=1.05µs min=0s med=0s max=2.15ms p(95)=0s p(99)=0s p(99.99)=1.53ms count=3000
http_req_duration..............: avg=228.72µs min=186.45µs med=219.13µs max=6.29ms p(95)=258.79µs p(99)=305.58µs p(99.99)=5.67ms count=3000
{ expected_response:true }...: avg=228.72µs min=186.45µs med=219.13µs max=6.29ms p(95)=258.79µs p(99)=305.58µs p(99.99)=5.67ms count=3000
✓ { gzip:yes }.................: avg=228.72µs min=186.45µs med=219.13µs max=6.29ms p(95)=258.79µs p(99)=305.58µs p(99.99)=5.67ms count=3000
http_req_failed................: 0.00% ✓ 0 ✗ 3000
http_req_receiving.............: avg=21.41µs min=14.37µs med=20.19µs max=136.93µs p(95)=27.72µs p(99)=36.52µs p(99.99)=130.43µs count=3000
http_req_sending...............: avg=60.9µs min=25.83µs med=33.25µs max=317.2µs p(95)=208.82µs p(99)=226.07µs p(99.99)=311.65µs count=3000
http_req_tls_handshaking.......: avg=4.19µs min=0s med=0s max=1.63ms p(95)=0s p(99)=0s p(99.99)=1.37ms count=3000
http_req_waiting...............: avg=146.39µs min=0s med=162.22µs max=6.22ms p(95)=184.61µs p(99)=229.52µs p(99.99)=5.59ms count=3000
http_reqs......................: 3000 99.994163/s
iteration_duration.............: avg=302.56µs min=245.1µs med=284.47µs max=6.38ms p(95)=348.05µs p(99)=462.9µs p(99.99)=5.82ms count=3000
iterations.....................: 3000 99.994163/s
vus............................: 20 min=20 max=20
vus_max........................: 20 min=20 max=20Process finished (33.45 seconds)
psrecord recorded cpu and memory usage on 4x core/8x thread Intel i7 4790K server. Update: now I've figured out how to insert the psrecord data into InfluxDB to be displayed via Grafana as well.
For k6 run itself
For Nginx web server tested
User 25, 50, 100, 0 concurrency 4 stage benchmarks without random sleep with psrecord. Where --duration $((TIME*5+30)) is the time multiple by the 4+1 stage runs + 30s defined in benchmark.js
TIME=30
DOMAIN=https://yourdomain.com/
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-no-sleep-nginx.log --plot plot-user-no-sleep-nginx.png &
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e STAGETIME=${TIME}s -e URL=$DOMAIN --no-usage-report --out json=summary-raw.gz benchmark.js" --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-no-sleep.log --plot plot-user-no-sleep.png
TIME=30
DOMAIN=https://yourdomain.com/
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-no-sleep-nginx.log --plot plot-user-no-sleep-nginx.png &
psrecord "taskset -c 0-3 k6 run -e STAGETIME=${TIME}s -e URL=$DOMAIN --no-usage-report --out json=summary-raw.gz benchmark.js" --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-no-sleep.log --plot plot-user-no-sleep.png
Starting up command 'taskset -c 0-3 k6 run -e STAGETIME=30s -e URL=https://domain1.com/ --no-usage-report --out json=summary-raw.gz benchmark.js' and attaching to process
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
execution: local
script: benchmark.js
output: json (summary-raw.gz)
scenarios: (100.00%) 1 scenario, 10 max VUs, 2m30s max duration (incl. graceful stop):
* default: Up to 10 looping VUs for 2m0s over 4 stages (gracefulRampDown: 30s, gracefulStop: 30s)
running (2m00.0s), 00/10 VUs, 372061 complete and 0 interrupted iterations
default ✓ [======================================] 00/10 VUs 2m0s
█ main index page
✓ is status 200
checks.........................: 100.00% ✓ 372061 ✗ 0
data_received..................: 881 MB 7.3 MB/s
data_sent......................: 15 MB 121 kB/s
group_duration.................: avg=716.71µs min=202.28µs med=366.9µs max=27.68ms p(95)=1.98ms p(99)=7.51ms p(99.99)=18.26ms count=372061
http_req_blocked...............: avg=295ns min=109ns med=218ns max=3.58ms p(95)=296ns p(99)=338ns p(99.99)=9.46µs count=372061
http_req_connecting............: avg=5ns min=0s med=0s max=871.71µs p(95)=0s p(99)=0s p(99.99)=0s count=372061
http_req_duration..............: avg=660.38µs min=159.34µs med=309.89µs max=36.12ms p(95)=1.9ms p(99)=7.39ms p(99.99)=21.27ms count=372061
{ expected_response:true }...: avg=660.38µs min=159.34µs med=309.89µs max=36.12ms p(95)=1.9ms p(99)=7.39ms p(99.99)=21.27ms count=372061
✓ { gzip:yes }.................: avg=660.38µs min=159.34µs med=309.89µs max=36.12ms p(95)=1.9ms p(99)=7.39ms p(99.99)=21.27ms count=372061
http_req_failed................: 0.00% ✓ 0 ✗ 372061
http_req_receiving.............: avg=129.3µs min=8.84µs med=19.65µs max=25.19ms p(95)=699.68µs p(99)=1.29ms p(99.99)=12.77ms count=372061
http_req_sending...............: avg=49.92µs min=16.27µs med=28.76µs max=22.97ms p(95)=182.66µs p(99)=248.71µs p(99.99)=11.58ms count=372061
http_req_tls_handshaking.......: avg=45ns min=0s med=0s max=3.41ms p(95)=0s p(99)=0s p(99.99)=0s count=372061
http_req_waiting...............: avg=481.15µs min=0s med=241.92µs max=27.27ms p(95)=1.09ms p(99)=6.73ms p(99.99)=15.53ms count=372061
http_reqs......................: 372061 3100.424592/s
iteration_duration.............: avg=1.28ms min=614.18µs med=901.1µs max=33.76ms p(95)=2.83ms p(99)=8.93ms p(99.99)=22.37ms count=372061
iterations.....................: 372061 3100.424592/s
vus............................: 1 min=1 max=10
vus_max........................: 10 min=10 max=10 Process finished (121.11 seconds)
Using jq to filter summary-raw.gz summary log for the 4 stages via the tags for each stage.
pzcat summary-raw.gz | jq -r 'select(.data.tags | .stage == "3")'
pzcat summary-raw.gz | jq -r 'select(.data.tags | .stage == "2")'
pzcat summary-raw.gz | jq -r 'select(.data.tags | .stage == "1")'
pzcat summary-raw.gz | jq -r 'select(.data.tags | .stage == "0")'
To filter for http_req_duration entries for tag for expected_response = true
pzcat summary-raw.gz | jq -c 'select(.metric == "http_req_duration" and .data.tags.expected_response == "true")'
To filter for http_req_duration entries that have status >= 200
pzcat summary-raw.gz | jq '. | select(.type=="Point" and .metric == "http_req_duration" and .data.tags.status >= "200")'
To average the value for http_req_duration entries that have status >= 200
pzcat summary-raw.gz | jq '. | select(.type=="Point" and .metric == "http_req_duration" and .data.tags.status >= "200") | .data.value' | jq -s 'add/length'
To min the value for http_req_duration entries that have status >= 200
pzcat summary-raw.gz | jq '. | select(.type=="Point" and .metric == "http_req_duration" and .data.tags.status >= "200") | .data.value' | jq -s min
To max the value for http_req_duration entries that have status >= 200
pzcat summary-raw.gz | jq '. | select(.type=="Point" and .metric == "http_req_duration" and .data.tags.status >= "200") | .data.value' | jq -s max
psrecord recorded cpu and memory usage on 4x core/8x thread Intel i7 4790K server. Update: now I've figured out how to insert the psrecord data into InfluxDB to be displayed via Grafana as well.
For k6 run itself
For Nginx web server tested
User 25, 50, 100, 0 concurrency 4 stage benchmarks with random sleep between 1 to 5 seconds with psrecord in benchmark2.js for export default function (). Where --duration $((TIME*5+30)) is the time multiple by the 4+1 stage runs + 30s defined in benchmark2.js
const sleepMin = 1;
const sleepMax = 5;and group('main index page', function ()
sleep(randomIntBetween(sleepMin, sleepMax));
TIME=30
DOMAIN=https://yourdomain.com/
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark2.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-nginx.log --plot plot-user-nginx.png &
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e STAGETIME=${TIME}s -e URL=$DOMAIN --no-usage-report --out json=summary-raw2.gz benchmark2.js" --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user.log --plot plot-user.png
TIME=30
DOMAIN=https://yourdomain.com/
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e STAGETIME=${TIME}s -e URL=$DOMAIN --no-usage-report --out json=summary-raw2.gz benchmark2.js" --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-nginx.log --plot plot-user.png
Starting up command 'taskset -c 0-3 k6 run -e STAGETIME=30s -e URL=https://domain1.com/ --no-usage-report --out json=summary-raw2.gz benchmark2.js' and attaching to process
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
execution: local
script: benchmark2.js
output: json (summary-raw2.gz)
scenarios: (100.00%) 1 scenario, 100 max VUs, 2m30s max duration (incl. graceful stop):
* default: Up to 100 looping VUs for 2m0s over 4 stages (gracefulRampDown: 30s, gracefulStop: 30s)
running (2m01.3s), 000/100 VUs, 1839 complete and 0 interrupted iterations
default ✓ [======================================] 000/100 VUs 2m0s
█ main index page
✓ is status 200
checks.........................: 100.00% ✓ 1839 ✗ 0
data_received..................: 4.5 MB 37 kB/s
data_sent......................: 142 kB 1.2 kB/s
group_duration.................: avg=2.94s min=1s med=3s max=5.01s p(95)=5s p(99)=5s p(99.99)=5.01s count=1839
http_req_blocked...............: avg=62.14µs min=199ns med=312ns max=10.03ms p(95)=822.43µs p(99)=1.07ms p(99.99)=9.88ms count=1839
http_req_connecting............: avg=16.21µs min=0s med=0s max=8.95ms p(95)=70.09µs p(99)=93.41µs p(99.99)=8.83ms count=1839
http_req_duration..............: avg=269.03µs min=188.43µs med=245.08µs max=11.23ms p(95)=352.24µs p(99)=479.78µs p(99.99)=9.31ms count=1839
{ expected_response:true }...: avg=269.03µs min=188.43µs med=245.08µs max=11.23ms p(95)=352.24µs p(99)=479.78µs p(99.99)=9.31ms count=1839
✓ { gzip:yes }.................: avg=269.03µs min=188.43µs med=245.08µs max=11.23ms p(95)=352.24µs p(99)=479.78µs p(99.99)=9.31ms count=1839
http_req_failed................: 0.00% ✓ 0 ✗ 1839
http_req_receiving.............: avg=27.11µs min=11.11µs med=22.3µs max=178.46µs p(95)=39.61µs p(99)=115.98µs p(99.99)=177.74µs count=1839
http_req_sending...............: avg=68.49µs min=29.51µs med=40.3µs max=399.02µs p(95)=237.77µs p(99)=303.61µs p(99.99)=398.04µs count=1839
http_req_tls_handshaking.......: avg=40.37µs min=0s med=0s max=1.22ms p(95)=664.3µs p(99)=843.76µs p(99.99)=1.19ms count=1839
http_req_waiting...............: avg=173.42µs min=0s med=175.37µs max=11.16ms p(95)=239.56µs p(99)=342.52µs p(99.99)=9.24ms count=1839
http_reqs......................: 1839 15.15872/s
iteration_duration.............: avg=2.94s min=1s med=3s max=5.01s p(95)=5s p(99)=5s p(99.99)=5.01s count=1839
iterations.....................: 1839 15.15872/s
vus............................: 3 min=1 max=100
vus_max........................: 100 min=100 max=100Process finished (121.99 seconds)
psrecord recorded cpu and memory usage on 4x core/8x thread Intel i7 4790K server. Update: now I've figured out how to insert the psrecord data into InfluxDB to be displayed via Grafana as well.
For k6 run itself
For Nginx web server tested
Send results to InfluxDB database using --out influxdb=http://localhost:8186/k6 where --duration $((TIME*5+30)) is the time multiple by the 4+1 stage runs + 30s defined in benchmark.js.
TIME=60
DOMAIN=https://yourdomain.com/
export K6_INFLUXDB_USERNAME=
export K6_INFLUXDB_PASSWORD=
export K6_INFLUXDB_PUSH_INTERVAL=2s
export K6_INFLUXDB_CONCURRENT_WRITES=$(nproc)
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark2.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-nginx.log --plot plot-user-nginx.png &
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e STAGETIME=${TIME}s -e URL=$DOMAIN --no-usage-report --out influxdb=http://localhost:8186/k6 benchmark.js" --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-no-sleep.log --plot plot-user-no-sleep.png
IME=60
DOMAIN=https://yourdomain.com/
export K6_INFLUXDB_USERNAME=
export K6_INFLUXDB_PASSWORD=
export K6_INFLUXDB_PUSH_INTERVAL=2s
export K6_INFLUXDB_CONCURRENT_WRITES=$(nproc)
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark2.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-nginx.log --plot plot-user-nginx.png &
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e STAGETIME=${TIME}s -e URL=$DOMAIN --no-usage-report --out influxdb=http://localhost:8186/k6 benchmark.js" --include-children --interval 0.1 --duration $((TIME*5+30)) --log psrecord-user-no-sleep.log --plot plot-user-no-sleep.png
Starting up command 'taskset -c 0-3 k6 run -e STAGETIME=60s -e URL=https://domain1.com/ --no-usage-report --out influxdb=http://localhost:8186/k6 benchmark.js' and attaching to process
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
execution: local
script: benchmark.js
output: InfluxDBv1 (http://localhost:8186)
scenarios: (100.00%) 1 scenario, 10 max VUs, 4m30s max duration (incl. graceful stop):
* default: Up to 10 looping VUs for 4m0s over 4 stages (gracefulRampDown: 30s, gracefulStop: 30s)
running (4m00.0s), 00/10 VUs, 648722 complete and 0 interrupted iterations
default ✓ [======================================] 00/10 VUs 4m0s
█ main index page
✓ is status 200
checks.........................: 100.00% ✓ 648722 ✗ 0
data_received..................: 1.5 GB 6.4 MB/s
data_sent......................: 25 MB 105 kB/s
group_duration.................: avg=854.19µs min=200µs med=412.45µs max=41.13ms p(95)=2.76ms p(99)=6.53ms p(99.99)=23.86ms count=648722
http_req_blocked...............: avg=277ns min=101ns med=211ns max=4.44ms p(95)=299ns p(99)=361ns p(99.99)=9.33µs count=648722
http_req_connecting............: avg=7ns min=0s med=0s max=2.29ms p(95)=0s p(99)=0s p(99.99)=0s count=648722
http_req_duration..............: avg=797.72µs min=156.47µs med=356.46µs max=62.08ms p(95)=2.68ms p(99)=6.3ms p(99.99)=25.44ms count=648722
{ expected_response:true }...: avg=797.72µs min=156.47µs med=356.46µs max=62.08ms p(95)=2.68ms p(99)=6.3ms p(99.99)=25.44ms count=648722
✓ { gzip:yes }.................: avg=797.72µs min=156.47µs med=356.46µs max=62.08ms p(95)=2.68ms p(99)=6.3ms p(99.99)=25.44ms count=648722
http_req_failed................: 0.00% ✓ 0 ✗ 648722
http_req_receiving.............: avg=262.93µs min=8.24µs med=20.91µs max=40.73ms p(95)=1.23ms p(99)=2.78ms p(99.99)=19.49ms count=648722
http_req_sending...............: avg=63µs min=15.17µs med=29µs max=33.02ms p(95)=198.24µs p(99)=437.99µs p(99.99)=13.72ms count=648722
http_req_tls_handshaking.......: avg=18ns min=0s med=0s max=1.58ms p(95)=0s p(99)=0s p(99.99)=0s count=648722
http_req_waiting...............: avg=471.78µs min=0s med=260.77µs max=32.3ms p(95)=1.25ms p(99)=3.81ms p(99.99)=16.57ms count=648722
http_reqs......................: 648722 2702.989389/s
iteration_duration.............: avg=1.47ms min=622.42µs med=984.91µs max=51.69ms p(95)=3.71ms p(99)=9.34ms p(99.99)=28.56ms count=648722
iterations.....................: 648722 2702.989389/s
vus............................: 1 min=1 max=10
vus_max........................: 10 min=10 max=10 Process finished (241.45 seconds)
InfluxDB + Grafana
Run k6 ramping-vus executor scenario and send results to InfluxDB database using --out influxdb=http://localhost:8186/k6 where --duration $((TIME*5+90)) is the time multiple by the 4+1 stage runs + 30s gracefulStop + 60s gracefulRampDown and variables VU=0, STAGEVU1=100, STAGEVU2=200, STAGEVU3=300, and STAGEVU4=0 are for the four stages defined in benchmark-scenarios-multi.js. Here VU=0 defines the starting VU number for ramping VUs in 4 stages from 0 to 100, 100 to 200, 200 to 300 and back down from 300 to 0.
TIME=60
VU=0
STAGEVU1=100
STAGEVU2=200
STAGEVU3=300
STAGEVU4=0
DOMAIN=https://yourdomain.com/
export K6_INFLUXDB_USERNAME=
export K6_INFLUXDB_PASSWORD=
export K6_INFLUXDB_PUSH_INTERVAL=2s
export K6_INFLUXDB_CONCURRENT_WRITES=$(nproc)
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark2.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+90)) --log psrecord-ramping-vus-nginx.log --plot plot-ramping-vus-nginx.png &
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e USERS=${VU} -e STAGETIME=${TIME}s -e STAGE_VU1=${STAGEVU1} -e STAGE_VU2=${STAGEVU2} -e STAGE_VU3=${STAGEVU3} -e STAGE_VU4=${STAGEVU4} -e URL=$DOMAIN --no-usage-report --out influxdb=http://localhost:8186/k6 benchmark-scenarios-multi.js" --include-children --interval 0.1 --duration $((TIME*5+90)) --log psrecord-ramping-vus.log --plot plot-ramping-vus.png
TIME=60
VU=0
STAGEVU1=100
STAGEVU2=200
STAGEVU3=300
STAGEVU4=0
DOMAIN=https://yourdomain.com/
export K6_INFLUXDB_USERNAME=
export K6_INFLUXDB_PASSWORD=
export K6_INFLUXDB_PUSH_INTERVAL=2s
export K6_INFLUXDB_CONCURRENT_WRITES=$(nproc)
# gather nginx resource usage via psrecord in background
# get Nginx MainPID value using Centmin Mod cminfo service-info tool
spid=$(cminfo service-info nginx | jq -r '.MainPID')
# set duration to 180 seconds as benchmark2.js uses 4x 30s stages + 30s = 2 1/2 min run time
psrecord $spid --include-children --interval 0.1 --duration $((TIME*5+90)) --log psrecord-ramping-vus-nginx.log --plot plot-ramping-vus-nginx.png &
# run k6 initiated via psrecord
psrecord "taskset -c 0-3 k6 run -e USERS=${VU} -e STAGETIME=${TIME}s -e STAGE_VU1=${STAGEVU1} -e STAGE_VU2=${STAGEVU2} -e STAGE_VU3=${STAGEVU3} -e STAGE_VU4=${STAGEVU4} -e URL=$DOMAIN --no-usage-report --out influxdb=http://localhost:8186/k6 benchmark-scenarios-multi.js" --include-children --interval 0.1 --duration $((TIME*5+90)) --log psrecord-ramping-vus.log --plot plot-ramping-vus.png
Starting up command 'taskset -c 0-3 k6 run -e USERS=0 -e STAGETIME=60s -e STAGE_VU1=100 -e STAGE_VU2=200 -e STAGE_VU3=300 -e STAGE_VU4=0 -e URL=https://domain1.com/ --no-usage-report --out influxdb=http://localhost:8186/k6 benchmark-scenarios-multi.js' and attaching to process
/\ |‾‾| /‾‾/ /‾‾/
/\ / \ | |/ / / /
/ \/ \ | ( / ‾‾\
/ \ | |\ \ | (‾) |
/ __________ \ |__| \__\ \_____/ .io
execution: local
script: benchmark-scenarios-multi.js
output: InfluxDBv1 (http://localhost:8186)
scenarios: (100.00%) 1 scenario, 300 max VUs, 4m0s max duration (incl. graceful stop):
* ramping_vus: Up to 300 looping VUs for 4m0s over 4 stages (gracefulRampDown: 0s, gracefulStop: 30s)
running (4m00.0s), 000/300 VUs, 861620 complete and 300 interrupted iterations
ramping_vus ✓ [======================================] 001/300 VUs 4m0s
INFO[0242] [k6-reporter v2.3.0] Generating HTML summary report source=console
█ main index page
✓ is status 200
checks.........................: 100.00% ✓ 861623 ✗ 0
data_received..................: 2.0 GB 8.5 MB/s
data_sent......................: 34 MB 141 kB/s
duration_in_seconds............: avg=0.037306 min=0.000168 med=0.03206 max=0.469338 p(95)=0.09501 p(99)=0.131728 p(99.99)=0.231943 count=861621
group_duration.................: avg=40.02ms min=218.47µs med=33.4ms max=469.43ms p(95)=103.61ms p(99)=160.33ms p(99.99)=342.84ms count=861620
http_req_blocked...............: avg=20.5µs min=113ns med=250ns max=206.61ms p(95)=367ns p(99)=510ns p(99.99)=76.69ms count=861630
http_req_connecting............: avg=10.14µs min=0s med=0s max=113.78ms p(95)=0s p(99)=0s p(99.99)=37.89ms count=861630
http_req_duration..............: avg=37.3ms min=168.32µs med=32.05ms max=469.33ms p(95)=95.01ms p(99)=131.72ms p(99.99)=231.94ms count=861630
{ expected_response:true }...: avg=37.3ms min=168.32µs med=32.05ms max=469.33ms p(95)=95.01ms p(99)=131.72ms p(99.99)=231.94ms count=861630
✓ { gzip:yes }.................: avg=37.3ms min=168.32µs med=32.05ms max=469.33ms p(95)=95.01ms p(99)=131.72ms p(99.99)=231.94ms count=861630
http_req_failed................: 0.00% ✓ 0 ✗ 861630
http_req_receiving.............: avg=10.92ms min=7.93µs med=6.63ms max=206.57ms p(95)=34.43ms p(99)=61.85ms p(99.99)=162.57ms count=861630
http_req_sending...............: avg=306.4µs min=18.73µs med=31.84µs max=387.88ms p(95)=110.79µs p(99)=8.35ms p(99.99)=96.93ms count=861630
http_req_tls_handshaking.......: avg=9.98µs min=0s med=0s max=163.72ms p(95)=0s p(99)=0s p(99.99)=37.37ms count=861630
http_req_waiting...............: avg=26.07ms min=0s med=23.15ms max=219.88ms p(95)=66.75ms p(99)=91.05ms p(99.99)=158.73ms count=861630
http_reqs......................: 861630 3590.086802/s
iteration_duration.............: avg=41.56ms min=612.03µs med=34.47ms max=592.75ms p(95)=106.55ms p(99)=167.67ms p(99.99)=403.28ms count=861620
iterations.....................: 861620 3590.045135/s
vus............................: 1 min=1 max=299
vus_max........................: 300 min=300 max=300 Process finished (242.16 seconds)
psrecord recorded cpu and memory usage on 4x core/8x thread Intel i7 4790K server. Update: now I've figured out how to insert the psrecord data into InfluxDB to be displayed via Grafana as well.
For k6 run itself
For Nginx web server tested
The recorded results in InfluxDB Grafana instance
k6 HTML report generated via forked k6 reporter
The recorded results in InfluxDB Grafana instance
