SPLK-3002 Implementing Services

Implementing Services Detailed Explanation

1. From Service Design to Deployment

Once you’ve carefully planned a service—identified what it monitors, chosen KPIs, and defined dependencies—it’s time to put the plan into action in ITSI.

This process is called service implementation, and it’s done using a special tool in ITSI called the Service Editor.

2. Steps in Service Implementation

Here’s a simple, step-by-step process to implement a service in ITSI:

Step 1: Create a New Service

• In the ITSI interface, go to the Service Analyzer or Service Configuration page.

• Click “Create New Service.”

• Give the service a clear, meaningful name like:

  • “Payment Gateway”

  • “Internal Email Service”

  • “Customer Login Portal”

You can also:

• Add a description

• Assign a team owner (if already defined)

Step 2: Add KPIs to the Service

Every service needs KPIs to measure how healthy it is.

You can:

• Create new KPIs from scratch by building new base searches

• Select existing KPIs from a template or library

For each KPI, define:

• The base search

• Split-by fields (if needed)

• Thresholds

• Importance (weight in service health calculation)

Step 3: Set Thresholds and Time Policies

Thresholds tell ITSI how to judge KPI values:

• Normal

• Warning

• Critical

There are three types of thresholds you can use (explained more in the next section).

You can also apply Time Policies to make your thresholds behave differently at different times of day (e.g., relaxed at night).

Step 4: Configure Dependencies

If this service depends on other services, you can define that relationship:

• Assign parent-child structure

• This means that if a child service goes critical, the parent service health score is also affected

Example:

The “Checkout Service” depends on “Database Service” and “Inventory Service”.

Dependencies help you understand the chain of impact in your IT environment.

Step 5: Assign to a Team and Configure Access

Each service should be managed by a team. In this step:

• Assign the service to the correct ITSI team (e.g., “Database Ops”, “App Dev”)

• Configure who can view or edit the service

• Set access control based on user roles and responsibilities

This keeps your monitoring environment organized, secure, and team-driven.

3. Threshold Management Options

When configuring KPIs in a service, you can choose from different types of threshold logic depending on the metric’s behavior.

a. Static Thresholds (Absolute Value)

• Simple and fixed

• Example: Error rate > 100 is “Critical”

Best for:

• Metrics with consistent behavior and well-known limits

b. Dynamic Thresholds (Percentage or Trending)

• Based on trends or percentage changes

• Example: If CPU usage increases 30% in 5 minutes, mark as “Warning”

Best for:

• Systems that vary naturally but still show warning signs when they change too quickly

c. Machine Learning-Based Thresholds (Anomaly Detection)

• Uses historical data to learn “normal” patterns

• Detects deviations from the learned baseline

• Works even if you don’t know exact limits

Best for:

• Complex environments

• KPIs with unpredictable behavior

You can set:

• Sensitivity level

• Learning period

• Re-training frequency

4. Time Policies – Why and How to Use Them

Time Policies let you apply different thresholds at different times of day. This is useful when system behavior naturally changes depending on the hour.

Example:

• At night, user traffic is low → Allow higher error tolerance

• During business hours, strict error limits are needed

Use Time Policies to:

• Reduce false alerts during maintenance windows

• Avoid unnecessary escalations

• Improve alert accuracy by adapting to daily patterns

You can define time policies using a calendar-style interface or custom time ranges.

Summary: What to Remember About Implementing Services

• Service implementation is the hands-on process of building your designed services in ITSI.

• You use the Service Editor to create services, assign KPIs, set thresholds, define dependencies, and configure access.

• ITSI gives you flexible options for threshold logic: static, dynamic, or machine learning-based.

• Use Time Policies to adjust KPI sensitivity based on business hours or known activity patterns.

Implementing Services (Additional Content)

1. Using Service Templates During Implementation

While service templates are introduced during service design, it's important to understand how they function at the implementation stage to accelerate service creation.

Applying Templates in Practice:

• When creating a new service in ITSI, you can select a pre-configured template from the dropdown in the Service Editor.

• The template provides:

  • Predefined KPIs (with base searches and thresholds)

  • Time policies

  • Health scoring logic

This ensures consistency across services and reduces setup time.

Benefits of Using Templates During Implementation:

• Rapid onboarding of new services

• Reduced human error in KPI setup

• Easy bulk updates if the template is modified later

Tip: Templates can be customized per service after application, but the core structure remains intact.

2. Health Score Calculation and KPI Importance

Every service in ITSI has a Health Score, typically a number between 0–100, reflecting the overall health of the service.

How It’s Calculated:

• The service health score is based on the status of all its KPIs.

• Each KPI contributes to the health score based on:

  • Its severity level (Normal, Warning, Critical)

  • Its importance weight (assigned during configuration)

Example:

• You may have three KPIs in a service:

  • CPU Load (weight 50%)

  • Disk Space (weight 30%)

  • Network Errors (weight 20%)

If CPU Load is Critical while others are Normal, the overall score will reflect that the most important metric is failing.

Tip: Carefully assign importance to each KPI during setup to align with business risk and operational impact.

3. Post-Implementation Verification and Adjustment

Once a service is configured and KPIs are in place, it’s essential to validate the implementation before relying on it in production.

Recommended Steps:

• Push sample/test data to see how KPIs behave.

• Use the Deep Dive tool to:

  • Visualize KPI behavior over time

  • Identify whether thresholds are too strict or too lenient

  • Detect missing data or unexpected status transitions

• Optionally use:

  • The KPI Trend Analysis dashboard

  • ITSI Health Check to ensure data integrity and proper indexing

Common Adjustments Post-Implementation:

• Tuning KPI thresholds to reduce false positives

• Changing importance weights based on observed impact

• Modifying base searches for better performance or relevance

Goal: Ensure that health scores and alerts accurately represent real-world service conditions.

Summary

• Templates can significantly streamline the implementation process and enforce best practices.

• Health scores depend on the behavior and weighting of KPIs—making it critical to assign meaningful importance levels.

• After service deployment, run validation checks using tools like Deep Dive or test data to confirm that KPIs behave as expected and thresholds are effective.