To run the performance test you currently have 2 options:
docker-compose up
This creates a mongodb instance, a catamel instance and a catanie instance.
It should return a time to ingest a number of datasets (currently 100)
scicatingest_1 | promise took 240 ms
This test the API server by running a set of API calls which simulate groups of different users
-
Install locust tool
Follow the installation description at https://docs.locust.io/en/stable/installation.html . This will provide a new command "locust" . Adjust your PATh if needed.
-
Setup API server
Any existing API server (catamel) instance with a connected DB backend will do. This can be your locally running development instance, a docker based catamel instance like above, or a production system.
Assuming your API server is running locally you simply do the following.
cd locust
locust --host=http://localhost:3000 WebsiteUser BeamlineIngestor
The command will look for the file locustfile.py in your current working directory. This file defines which type of users exist and what API calls should be run. An example file is in the folder locust/locustfile.py. You then open http://localhost:8089 and enter the number of users to simulate as well as the "hatch rate", i.e. how fast users are spawned per second. The browser then presents resulting statistics both in tabular and chart form. Below you find two examples with 10 and 100 users respectively:
The following command will run tests for 60 seconds with 100 users and a hatch rate of 5/s:
locust --host=http://localhost:3000 --no-web -c 100 -r 5 --run-time 60s WebsiteUser BeamlineIngestor
The final summary table looks like this:
Name # reqs # fails Avg Min Max | Median req/s
--------------------------------------------------------------------------------------------------------------------------------------------
GET API index 9 0(0.00%) 111 4 792 | 28 0.20
DELETE Delete old test datasets 1260 0(0.00%) 9 7 37 | 8 0.00
GET Facetquery for 5 facets with given ownerGroup 37 0(0.00%) 747 93 3841 | 360 1.00
GET Fullquery for 30 datasets in given ownerGroup 40 0(0.00%) 1477 223 4291 | 400 0.80
POST Ingest raw dataset 4132 0(0.00%) 117 9 8038 | 72 148.40
POST Login as ingestor 102 0(0.00%) 4194 85 6703 | 4900 0.00
POST Logout as ingestor 100 0(0.00%) 138 49 272 | 130 0.00
GET Query for 100 datasets in given ownerGroup 44 0(0.00%) 583 26 3358 | 120 1.50
GET Query for 2 datasets in given ownerGroup 204 0(0.00%) 683 13 4502 | 110 6.90
GET Query for old test datasets 1 0(0.00%) 104 104 104 | 100 0.00
--------------------------------------------------------------------------------------------------------------------------------------------
Total 5929 0(0.00%) 158.80
Percentage of the requests completed within given times
Name # reqs 50% 66% 75% 80% 90% 95% 98% 99% 100%
--------------------------------------------------------------------------------------------------------------------------------------------
GET API index 9 28 31 35 44 790 790 790 790 790
DELETE Delete old test datasets 1260 8 9 9 9 11 12 13 17 37
GET Facetquery for 5 facets with given ownerGroup 37 360 470 580 610 2700 3400 3800 3800 3800
GET Fullquery for 30 datasets in given ownerGroup 40 400 1900 3200 3400 3900 4100 4300 4300 4300
POST Ingest raw dataset 4132 72 89 110 130 220 300 380 600 8000
POST Login as ingestor 102 4900 5000 5000 5100 5200 5900 6400 6500 6700
POST Logout as ingestor 100 130 160 180 190 220 240 250 270 270
GET Query for 100 datasets in given ownerGroup 44 120 160 190 250 2600 3300 3400 3400 3400
GET Query for 2 datasets in given ownerGroup 204 110 170 290 1600 3200 3400 3500 3600 4500
GET Query for old test datasets 1 100 100 100 100 100 100 100 100 100
--------------------------------------------------------------------------------------------------------------------------------------------
Total 5929 63 84 100 130 240 360 3300 4900 8000