Releases: AntsInMyEy3sJohnson/hazeltest
v0.14.0
Changes since v0.13.0
The state cleaners introduced in scope of #40 have been integrated with the status API. The data structures they have cleaned are accessible via the application's status endpoint, :8080/status
.
Corresponding image on DockerHub: antsinmyey3sjohnson/hazeltest:0.14.0
v0.13.0
Changes since v0.12.1
Introduce state cleaners for maps and queues.
Beyond the eviction flag found on map runners, the new state cleaners can also consider data structures not under their management. So, if a previous Hazeltest instance was run with a configuration that advised its runners to append that Hazeltest instance's unique ID to each data structure name, the state cleaners can clean such data structures, too.
Ultimately, this feature is useful in all cases where the load test operator wishes to make sure that each load test starts on a completely empty Hazelcast cluster.
See also #40.
Corresponding image on DockerHub: antsinmyey3sjohnson/hazeltest:0.13.0
v0.12.1
v0.12.0
Changes since v0.11.0
Introduce possibility to have map test runner's batch test loop sleep after executing one batch action.
Making batch test loops sleep after a batch action is useful for artificially slowing down the number of operations on the Hazelcast cluster under test in order to, for example, align the load more to what load in production is like.
Corresponding image on DockerHub: antsinmyey3sjohnson/hazeltest:0.12.0
v0.11.0
Changes since v0.10.1
Introduce more information to status endpoints.
Runners now report various failure events of their test loops to the status gatherer, which delivers it to the application's /status
endpoint, when queried. Users can build evaluation logic on top of these reports so as to automatically determine, for example, whether a load test was successful or has failed.
Corresponding image on DockerHub: antsinmyey3sjohnson/hazeltest:0.11.0