A Clojure library designed to generate and execute requests against a web service based on a Swagger specification.
There are several example files in the resources folder. In particular, check out the fully-annotated petstore-mixed.edn example.
Docker image names are based on the git revision. The following
sequence of commands will build a docker image named
requestinator:v1.2.3. It will also tag it latest.
git commit -am "Message for commit"
git tag v1.2.3
bin/build
Generate three indepdendent, random sequences of requests based on the
Petstore Sample Service, and save it to
/tmp/requestinator-test. Requests will be scheduled on average
twice a second for each of the three agents, and 60 seconds worth of
data will be generated.
lein run generate --destination file:///tmp/requestinator-test --params resources/petstore-mixed.edn
docker run -e AWS_ACCESS_KEY_ID -e AWS_SECRET_KEY requestinator generate --destination s3://com.cognitect.requestinator.test/readme-example --params resources/petstore-mixed.edn
This example assumes you have AWS_ACCESS_KEY and AWS_SECRET_KEY environment variables set to provide access to the com.cognitect.requestinator.test bucket.
lein run execute --source file:///tmp/requestinator-test --destination file:///tmp/requestinator-test/results --recorder-concurrency 3
docker run -e AWS_ACCESS_KEY_ID -e AWS_SECRET_KEY requestinator execute --source s3://com.cognitect.requestinator.test/readme-example --destination s3://com.cognitect.requestinator.test/readme-example/results --recorder-concurrency 3
lein run report --source file:///tmp/requestinator-test/results --destination file:///tmp/requestinator-test/reports
Open /tmp/requestinator-test/reports/main/html/index.html to view.
There is one row for each agent, in which triangles represent
requests, positioned horizontally according to the time they were
executed. The length of the solid rectangle indicates the duration.
Hover over a request to see its details in the lower pane. Click one
to "lock" it. Hold shift and scroll to zoom the time dimension in and
out.
Requestinator supports spreading execution of requests across multiple processes (and therefore multiple machines). This is accomplished by selecting a subset of agent groups (defined in the parameters file) and a start time for each executor process. Activity generation is unchanged and cannot currently be distributed.
Assuming the same generation as above, the following commands would set up two separate processes to begin execution at 12:34PM. Process 1 will run the agents in group "markov" and Process 2 will run the agents in group "uniform".
# Process 1
lein run execute --source file:///tmp/requestinator-test --group markov --destination file:///tmp/requestinator-test/results/markov --start-time 12:34 --recorder-concurrency 3
# Process 2
lein run execute --source file:///tmp/requestinator-test --group uniform --destination file:///tmp/requestinator-test/results/uniform --start-time 12:34 --recorder-concurrency 3
Note that the destinations must be separate directories, as each process will assume it has full ownership of the files in that directory.
A unified report can be generated from the two results by pass multiple --source switches to the report command:
lein run report --source file:///tmp/requestinator-test/results/markov --source file:///tmp/requestinator-test/results/uniform --destination file:///tmp/requestinator-test/reports
- Implement the rest of Swagger. Pretty good coverage at the moment,
but definitely some of the spec is not implemented. Partial list:
- Support for
requiredbeing false - Support for XML
- Overriding consumes
- Support for
- Implement the rest of JSON Pointer.
- Improve tests
Copyright © 2016 Cognitect, Inc.