AWS appconfig documentation change
Summary
Removed entire section about exporting experiment assignment data including logging configuration, file structure, and JSONL format
Security assessment
Removal of data collection documentation doesn't indicate a security fix. No vulnerability references or security advisories are mentioned in the change.
Diff
diff --git a/appconfig/latest/userguide/appconfig-experimentation-about-data-collection.md b/appconfig/latest/userguide/appconfig-experimentation-about-data-collection.md index a134188f1..f0f858224 100644 --- a//appconfig/latest/userguide/appconfig-experimentation-about-data-collection.md +++ b//appconfig/latest/userguide/appconfig-experimentation-about-data-collection.md @@ -15,26 +14,0 @@ For experiment results analysis such as conversion rates, user engagement, and f -**Exporting experiment assignment data** - -To capture which users are assigned to which treatments during an experiment run, configure AWS AppConfig Agent with the `EXPERIMENT_ASSIGNMENT_LOG_DESTINATION` option. When configured, the agent writes assignment logs to the specified destination. If this option is not set, no assignment logs are emitted. - -Currently, the agent supports writing logs to disk. Specify a file path as the destination value (for example, `file:/var/log/appconfig/experiments/`). - -Assignment logs are stored in the following directory structure: - - - BASE_DIR/ - <region>/ - <account-id>/ - <application-id>/ - <experiment-definition-id>/ - <run-number>_<unique-id>.jsonl - -Each log file contains one JSON record per line (JSONL format). Each record includes the timestamp, the treatment assignment, and the entity ID: - - - {"timestamp":"2026-01-01T12:00:00Z","treatmentKey":"__t1__","entityId":"user1"} - {"timestamp":"2026-01-01T12:00:01Z","treatmentKey":"__c__","entityId":"user2"} - -After logs are written to disk, you are responsible for exporting them to persistent storage such as Amazon S3 or a data warehouse for analysis. - -After you begin an experiment, verify that data collection works correctly for the selected audience. Ensure that tracking and logging are functioning as expected. Also, confirm that data is complete and available for analysis. Incomplete or inaccurate data can invalidate experiment results. -