Release Notes
Last updated
Was this helpful?
Last updated
Was this helpful?
Welcome to the Akamas 3.6 release notes! This update is another step in the journey to improve user experience and simplify the use of Akamas in many different contexts. Let's jump into the core improvements of this release.
We received a lot of feedback from customers who would like to interrupt study execution and resume it again at a later time to cope with shared performance environments or environment configuration freeze periods. For this reason, we have complemented the ability to stop a running study, already available in previous versions, to also resume it at any given time. This allows for greater flexibility in coping with schedule changes.
The is one of the most commonly used operators of workflows. It allows to actualize configuration file templates with the parameter values optimized by Akamas. Each technology comes with a set of parameters that are applied with a specific format, this information is already included in Optimization Packs and allows the file configurator operator to work out of the box. In some scenarios, user-defined practices or other configuration tools (e.g. helm charts, ansible playbooks) override the default way to set a parameter. To improve the experience in this scenario, the file configurator operator allows now to override the way a parameter is substituted in order to specify any custom format.
Offline studies can now start their exploration from any baseline configuration, even if it does not fulfill the constraints defined in the study. This helps to explore different parameter regions with respect to the one already used by the system and quickly compare them.
Live studies still need a baseline configuration that respects parameter constraints but can not start from baselines that do not completely fulfill goal constraints. This allows you to start optimizing applications that are also experiencing some response time or resource utilization spikes for a limited time.
Troubleshooting telemetry instance integration is an activity that is often performed during the initial study setup. To simplify this activity all telemetry instances are now automatically executed with the highest log level. You can now filter for different log levels directly in the UI. The UI will also highlight failed telemetry instances and direct you straight to the error message to speed up troubleshooting and get to the optimization faster.
The first revision of 3.6 includes many usability updates focused on reducing the effort required to apply Akamas and improving troubleshooting activities.
In particular, the upgrade of telemetry providers has been redesigned to remove the need to remove and recreate telemetry instances manually.
Offline studies now can also contain experiments with configurations that do not fulfill parameter constraints, even though they don't provide useful information for the optimization process, they can be used as a reference for quick comparison.
Exporting and importing a study now also extracts logs and the input of the optimization process. This allows extracting relevant study and troubleshooting information in a single archive that can be shared with customer support for improved interactions and kept as an offline backup of executed studies.
Loadrunner Enterprise support has been extended to versions 2023 R1 and 2024
Here you can find a list of other notable changes in this revision.
Added the capability to delete trials and experiments of live studies
Allow to change manual/automatic approval for stopped live studies
Allow users to download the entire dataset from the analysis table with a single click
Added the ability to filter trials in the analysis table by parameter values
Ability to download all study logs for support activities with a single click
Improve smoothness and responsiveness during a study run.
Show incompatible telemetry instances after a telemetry provider has been upgraded
Added the ability to select an aggregation for metrics defined as KPIs
Added "--installed" Flag to show installed optimization packs first in "akamas list optimization-pack" command
Added a force option when deleting a non-empty workspace
Improved error messages
Deleting a workflow caused logs of the study to be not available
Could not bootstrap experiments from a study whose system has been deleted
Only the first failed constraints were reported during the stability windowing evaluation
Exporting studies did not keep track of preferred experiments and custom tags
The optimization process failed if some data points were missing metrics required for the analysis.