* added cypress setup for fleet
* updated readme
* fixed types
* fix unused
* starting up fleet server in cypress
* added more fleet tests
* added package update test
* updated readme
* trying jenkins run
* trying jenkins run
* trying jenkins run
* fix types, cypress config
* example of mocked requests
* added getBySel
* fix test after refactor
* removed duplication
* fix tests
* added to buildkite
* updated tests
* starting agent with docker
* trying to fix ip address
* ifconfig
* ip address
* ip address
* ip address
* type fix
* remove extra logging
* fixed test
* fix buildkite
* cleanup
* cleanup
* using docker for fleet_server
* fix
* fix
* trying to fix
* update config
* test
* test
* test
* test
* revert changes
* test
* test
* static ip on linux
* docker version
* try again
* fixed review comments
* fixed types
Co-authored-by: Kibana Machine <42973632+kibanamachine@users.noreply.github.com>
Co-authored-by: Patryk Kopycinski <contact@patrykkopycinski.com>
Concurrency parameters were not in the correct place which made concurrency settings not in effect.
Moved it to to performance.sh step which also has parallelism in the same level.
Co-authored-by: Kibana Machine <42973632+kibanamachine@users.noreply.github.com>
* Re-enable APM E2E tests and allow server to shut down cleanly on failure
Calling `process.exit` in the test script made it so the FTR runner would not properly shut down the server, and cause other tests to fail because Kibana was left running on a port. Remove that and just throw
instead.
No changes were made to the tests, as I was unable to reproduce any failures locally. I'll try in CI and see if we can get anything to fail.
Fixes#115280.
This is a short-term solution to reduce our hit count in Buildkite due to the increase in bootstraps so we can move forward with enabling Buildkite across all PRs.
Adds auto and manual performance tests
If "ITERATION_COUNT_ENV" exist in build job:
- Auto Mode: Checks "ITERATION_COUNT_ENV",
- Manual Mode: Waits for user input,
use provided value to run "n" number of performance tests on parallel
- Sends slack notifications when pipeline finished
* Re-enable previously disabled APM E2E tests.
* Round to the nearest second in `getComparisonTypes` to avoid cases where a millisecond difference can change which results get shown.
* Simplify error count alert tests to test the "happy path" (#79284 exists in order to expand to more tests for rule editing and creation.)
* Wait for alert list API request to complete before clicking "Create rule" button when running the test to create a rule from the Stack Management UI.
I ran the e2e tests 100 times locally with no failures so I'm confident the flakiness has been addressed.
Fixes#114419.
Fixes#109205.