-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Cache TerraformTest methods #57
Conversation
Hey Marshall, thanks for this. Could you briefly describe a concrete use case? If we add this, we should add some text to the README explaining typical usage. |
Hi @ludoo, thanks for the recommendation. I made the appropriate changes to the PR description under the Use Cases section |
Marshall, sorry for the delay on this. Would you mind adding a section to the top-level README that explains the fixtures, and give a quick usage example? Then this is good to go, even though I'm still struggling to understand it in full (my problem, not yours). :) |
@ludoo no worries, I'm glad it wasn't merged given it gave me the time to figure out how to simplify and improve the PR. I made some big changes including removing the fixtures entirely and adding the caching feature internally to the This removes the need to use the fixtures in order to benefit from the caching feature. I also realized that the fixtures are pretty general and it's likely that users are going to want to customize the logic within the tftest fixtures anyways. I added a description and a usage snippet to the README. Let me know if there's anything else I should add. The idea is kind of confusing but I'm hoping the changes I just introduced should simplify it. Is there a particular aspect of it that you need more clarification about? I'd be happy to answer any questions you may have. |
I can never figure out who can actually see the merge button, if you cannot just let me know. |
The changes introduced incorporate PyTest fixtures that cache tftest method outputs. The outputs can be used for future pytest session invocation without users having to rerun potentially time-intensive Terraform commands. The fixtures originated from a repo I created called pytest-terra-fixt although I thought that this may be useful to add to the tftest package.
Fixtures
All fixtures below can be parametrized via (
@pytest.mark.parametrize()
)terra
:binary
: Path to binary (must end withterraform
or `terragrunt)tfdir
: Absolute or relative path tobasedir
basedir
: Base directory fortfdir
(defaults to cwd)env
: Environment variables to pass to the commandtftest.TerragruntTest
ortftest.TerraformTest
class depending on thebinary
parameterterra_setup
A: Dictionary of keyword arguments to passed to the
tftest.TerraformTest.setup()
methodB: Dictionary of terra directories and their respective keyword arguments to pass to the method mentioned above
terra
fixture's associated setup outputterra_plan
A: Dictionary of keyword arguments to passed to the
tftest.TerraformTest.plan()
methodB: Dictionary of terra directories and their respective keyword arguments to pass to the method mentioned above
terra
fixture's associated plan outputterra_apply
A: Dictionary of keyword arguments to passed to the
tftest.TerraformTest.apply()
methodB: Dictionary of terra directories and their respective keyword arguments to pass to the method mentioned above
terra
fixture's associated apply outputterra_output
A: Dictionary of keyword arguments to passed to the
tftest.TerraformTest.output()
methodB: Dictionary of terra directories and their respective keyword arguments to pass to the method mentioned above
terra
fixture's associated outputExamples
conftest.py
test_tf.py
Use Cases
The following are short and long versions of the possible use case
Brief use case
to have to wait for their Terraform setup logic to run every time
Long use case
Terraform modules can take a long time to spin up. If you're developing tests locally for your modules and the teardown logic or tftest setup method includes the
cleanup_on_exit=True
argument, then subsequent test invocations will have to run the entire Terraform deployment flow again in order for your tests to run. This may be ideal if the Terraform module source code changes in which you want the refreshed outputs to be used for testing assertions. Although if you want a fast feedback loop for writing assertions and the module code hasn't changed, then the tftest method fixtures may come in handy given its caching abilities. An alternative is just to comment out any teardown logic or setcleanup_on_exit=False
, although tftest will still have to call the requested Terraform command even though its associated output hasn't changed. This process is still slower compared to using the local caching feature.Notes
I'm more than happy to change any fixture names or general styling/formatting to be more coherent with the tftest package.