# README

Collisions test for the context key generator

Run run_test.sh after having read its sources. Note that it runs for 17 minutes on a 2018 MacBook Pro i7 quad-core and needs more than 3GiB of RAM.

Use generate_contexts.py to generate random entries in a file called random_contexts.csv Each line will have the format:

<metric_name>,<tag1> <tag2> ...

where the number of tags is random between 1 and 5. All values (metric name and tags) will be random UUID4.

Once this file has been generated, make sure there is no duplicate contexts in there:

cat random_contexts.csv|sort|uniq > random_sorted_uniq_contexts.csv

It will run for several minute. Finally, run the unit test:

$ pwd
datadog-agent/pkg/aggregator/ckey/tests
$ go test ./... -run TestCollisions

If an error happen, it means that it has detected a collision.

This test will use a lot of RAM (3GiB on my machine)