Skip to main content
Skip table of contents



Germain UX offers different types of SLAs (Service Level Agreements) that can be applied to various KPIs, whether they are preconfigured or custom, and whether they are related to business or technical metrics.

Video overview:

Here are the different types of SLAs available in Germain UX:

Fact-Based SLA

This type of SLA evaluates individual data points or facts. It allows for the evaluation of specific metrics against predefined thresholds or targets.

Statistical SLA

Statistical SLAs evaluate collections of data points or facts. Several types of statistical SLAs are available:

Aggregate SLA

It evaluates a specific period of time and compares it to a constant threshold. This type of SLA is useful for monitoring trends or performance over a specified time frame.

Color Percentile SLA

This SLA evaluates a period of time and compares the percentile of data points in the GREEN range to a constant threshold. It helps identify if performance is within an acceptable range.

Percent Change SLA

It evaluates a period of time and compares it to a prior period, measuring the percentage change. This SLA is useful for monitoring performance improvements or regressions over time.

  • “In Range” operator

“In Range” operator can be used with a Percent-change SLA and a Threshold of -/+30%


Example of an alert leverage in range operator:


Example of an alert leverage in range operator - Germain UX


Predictive SLA

This SLA creates predictions based on past data points or historical trends. It helps in forecasting future performance or identifying potential issues based on predictive analytics.

Smart Insights SLA

This SLA automatically identifies trends that are considered abnormal or noteworthy. It evaluates a business period of time (e.g., year-end, quarter-end, month-end, specific weekdays) and compares it to other similar business periods. Smart Insight SLAs provide alerts or notifications for significant deviations or insights.

These SLAs help organizations set performance benchmarks, monitor adherence to targets or thresholds, and gain insights into trends and predictive analytics. By utilizing different types of SLAs, Germain UX enables users to evaluate and track KPIs effectively, both in real-time and historically.

Examples of smart insights in Germain UX include predictions about disk usage exceeding a certain threshold, changes in user errors volume compared to previous periods, sales forecast accuracy predictions, and predictions of visitor increases in eCommerce by month-end.

Germain UX’ Smart Insight engine can be enabled for any of the KPIs that are active in your system. KPIs can be Technology, Process or User Behavior-specific.

In summary, Germain UX provides a range of SLA types, including fact-based and statistical SLAs, to monitor and evaluate KPIs. These SLAs enable organizations to set performance targets, identify trends, and gain predictive insights. The Smart Insight SLAs offer automated detection of abnormal trends, providing valuable information for proactive monitoring and decision-making.


  • Alert1: Disk Usage is predicted to exceed 90% in a week from now

  • Alert2: User Errors Volume has increased by 22.7% this monday compared with last monday

  • Alert3: Sales Forecast predicted to be 70% accurate by month-end

  • Alert4: eCommerce visitors predicted to be 40% higher by month-end


Smart Insights Summary Email that is sent once per day.

Filter Percent Change SLA with Baseline and Current Results

Germain now includes the capability to apply constraints to a percent change Service Level Agreement (SLA) based on baseline and current results.

A new configuration property called "Threshold Expression" has been introduced for statistical SLAs, which can support advanced constraints. It allows the utilization of the following variables:

  • Current: Represents the absolute value for the "current" period as defined by the SLA.

  • Baseline: Represents the absolute value for the "baseline" period as defined by the SLA.

  • Value: Represents the change between the baseline and current values, either absolute or in percentage, depending on the SLA configuration.

  • Threshold: Represents the threshold value defined by the SLA.

For example, if both the baseline and current counts are less than 20, even if the percentage change exceeds a configured threshold, no alert will be triggered.

(current < 20 && baseline < 20) || (value < threshold)


SLA evaluation for multiple metrics on a single data point

Before, only one metric (duration, value, etc) could be evaluated against the SLAs configured for a given data point. This feature allows Germain UX to support having SLAs for both multiple metrics (e.g. Inactive Duration and Active Duration, as well as Color threshold values for both).

Use Case

This client was only interested in identifying where 80 employees are spending their time while creating HR plans later handed over to HR brokers. Every minutes is significant to this client business, the reason why they needed Germain UX to analyze all this in real-time and provide actionable insights as to why and how time is spent.


As part of this enhancement, Germain UX now supports having SLAs for as many metrics as you like.


Here is an example a User Click SLA taking into account 3 separates criteria or thresholds: slowness, failure or large memory footprint.

User Click SLA - Germain UX

Thresholds are always evaluated in order, with the "worst" color ultimately being put onto the SLA (i.e., one threshold colors it RED, the other would color it ORANGE, the fact ends up with RED).

1st threshold uses expression to check if triggered or not:

1st threshold of User Click SLA - Germain UX

2nd threshold evaluates against KPIs default property (duration):

2nd threshold of User Click SLA - Germain UX

3rd threshold evaluates against non-default property:

3rd threshold of User Click SLA - Germain UX

Service: Analytics

Feature Availability: 8.6.0 or later

JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.