Release Notes

What's new in 3.1.3

This version brings several enhancements in the integration with Dynatrace and Prometheus, in particular in the context of optimizations based on Kubernetes and Java whose optimization packs have been reviewed and extended to better capture the details of these technologies and allow users to build simpler and more powerful optimization studies.

Other enhancements are focused on improving the user experience from the creation of studies to the exploration of the results.

Here is a comprehensive list of enhancements and bugfixes you can find in this version:

User Experience

  • The user is warned about the lack of telemetry instances when starting the study.

  • When running live optimizations, users can now manually approve configurations also for baseline and preset steps, not just the optimization step.

UI

  • Simplify the selection of baseline, best and user-starred trials in the metrics chart.

  • Easily share charts and tables using a link with embedded filters.

  • Improve reactivity of KPI management and study data refresh

  • Allow users to delete components via UI.

  • Simplify study navigation keeping the user selection in metrics and table filters.

Prometheus Provider

  • Review and improve metrics for Java, OpenJ9, and Kubernetes

  • Improve support for custom connection header to support custom installations (i.e. Prometheus derivations such as Cortex or custom proxies)

  • Support custom query duration via DURATION property in the telemetry instance

  • Allow users to specify custom filters in the components using only the %FILTERS% token in queries.

Dynatrace Provider

  • Review and improve metrics for Java, OpenJ9, and Kubernetes

  • Simplified component mapping for Kubernetes (Container, Job and Workload)

API

  • Entich metrics retrieval API to allow the specification of a time filter instead of a trial-based filter.

  • Introduce an API to retrieve information for all trials in a study.

Bugfixes

  • Allow users to delete Workflow by name instead of id via cli.

  • Avoid failing telemetry with multiple concurrent telemetry instances terminating at the same time.

  • The radar chart collapses when removing a KPI without issuing a reset.

  • Metrics chart failure when trying to show trials for which a window have not been computed.

  • Missing experiments in analysis table csv export.

  • Only optimized parameters are shown in the configuration drawer, instead of all applied parameters.

  • Allow filtering for study or services via cli in the akamas logs command

  • Multiple concurrent license usages are created for the same study upon study restart

  • A misleading error message (telemetry failure) is shown to the user when a workflow task fails

Last updated