Review and apply recommendations

Review and manage recommendations from the Adaptive Traces UI.

Get started

When you enter the Adaptive Traces application for the first time, you apply the recommendation to create your first probabilistic policy.

Probabilistic policies sample a certain percentage of traces and you can specify the sampling rate. In this case, you create a probabilistic policy sampling 5% of traces.

Navigate to Administration > Cost management > Traces cost management > Adaptive Traces.
Click Apply recommendation to create your first policy.
The following policy is created:
Name: Sample five percent of all traces (probabilistic)
Body:
JSON
```
{
  "sampling_percentage": 5
}
```
You can edit it and change its name as needed.

Note
If this policy is removed before another policy is created, you are re-prompted to begin with this policy the next time you use Adaptive Traces.

Review recommendations

From the Overview page, you can review any new recommendations or check the Recommendations History feed.

New recommendations are automatically generated based on regular analysis of your tracing data, helping you fine-tune your trace sampling and ensuring you only capture what’s critical and worthy of your attention.

The Recommendations History feed helps you with the following:

Understand which recommendations were applied and when, providing you with a record of changes made.
Correlate applied recommendations with changes in data ingestion volume, costs, or debugging efficiency.
Identify if a recently applied or dismissed recommendation is the cause of changes in your data.
Assess whether applied recommendations lead to the outcomes you expect.

You can filter for All, Applied, Dismissed, Stale.

Note
A stale recommendation is one that is outdated. For example, the recommendation engine runs, sees that you do not have a status code policy and creates a recommendation to create one. You do not apply it, but create one yourself. The recommendation engine runs again and sees that a status code policy exists and marks the unapplied recommendation as stale.
If you applied a recommendation, but later delete the policy that resulted from the recommendation, it still shows in the Recommendation History with status Applied. It is also re-generated as a new recommendation.