Platform observability and SLAs in MLOps - Mini Project: Build & Apply

Platform Observability and SLAs

📖 Scenario: You work as a DevOps engineer for a machine learning platform team. Your team wants to monitor the platform's health by tracking service uptime and response times. They also want to check if the platform meets the agreed Service Level Agreements (SLAs).SLAs require the platform to have at least 99% uptime and average response time below 200 milliseconds.

🎯 Goal: Build a simple Python script that stores platform metrics, sets SLA thresholds, calculates uptime and average response time, and prints whether the platform meets the SLAs.

📋 What You'll Learn

Create a dictionary with exact platform metrics data

Add SLA threshold variables for uptime and response time

Calculate uptime percentage and average response time using loops

Print the SLA compliance results exactly as specified

💡 Why This Matters

🌍 Real World

Monitoring platform health and ensuring it meets SLAs is critical for reliable machine learning services.

💼 Career

DevOps engineers and MLOps specialists use observability and SLA checks daily to maintain service quality.

1

Create platform metrics data

Create a dictionary called platform_metrics with these exact entries: 'uptime_minutes': [1440, 1430, 1420, 1440, 1435] and 'response_times_ms': [180, 210, 190, 170, 200].

MLOps

# Create the platform_metrics dictionary with uptime_minutes and response_times_ms lists
# Your code here

Need a hint?

Use a dictionary with two keys: 'uptime_minutes' and 'response_times_ms'. Each key should have a list of integers as values.

2

Add SLA threshold variables

Add two variables: sla_uptime_threshold set to 99.0 and sla_response_time_threshold set to 200.

MLOps

platform_metrics = {
    'uptime_minutes': [1440, 1430, 1420, 1440, 1435],
    'response_times_ms': [180, 210, 190, 170, 200]
}
# Add SLA threshold variables below
# Your code here

Need a hint?

Set sla_uptime_threshold to 99.0 (percent) and sla_response_time_threshold to 200 (milliseconds).

3

Calculate uptime percentage and average response time

Calculate the total possible uptime minutes as 1440 * 5. Calculate the actual uptime by summing platform_metrics['uptime_minutes']. Calculate uptime_percentage as (actual uptime / total possible uptime) * 100. Calculate average_response_time as the average of platform_metrics['response_times_ms']. Use for loops with variables minute and time to sum the lists.

MLOps

platform_metrics = {
    'uptime_minutes': [1440, 1430, 1420, 1440, 1435],
    'response_times_ms': [180, 210, 190, 170, 200]
}
sla_uptime_threshold = 99.0
sla_response_time_threshold = 200

# Calculate uptime_percentage and average_response_time below
# Your code here

Need a hint?

Use for loops to sum the uptime and response times. Then calculate percentages and averages.

4

Print SLA compliance results

Print two lines exactly as follows: print(f"Uptime meets SLA: {uptime_percentage >= sla_uptime_threshold}") and print(f"Response time meets SLA: {average_response_time <= sla_response_time_threshold}").

MLOps

platform_metrics = {
    'uptime_minutes': [1440, 1430, 1420, 1440, 1435],
    'response_times_ms': [180, 210, 190, 170, 200]
}
sla_uptime_threshold = 99.0
sla_response_time_threshold = 200

total_possible_uptime = 1440 * 5
actual_uptime = 0
for minute in platform_metrics['uptime_minutes']:
    actual_uptime += minute

uptime_percentage = (actual_uptime / total_possible_uptime) * 100

total_response_time = 0
for time in platform_metrics['response_times_ms']:
    total_response_time += time

average_response_time = total_response_time / len(platform_metrics['response_times_ms'])

# Print SLA compliance results below
# Your code here

Need a hint?

Use print statements with f-strings to show if uptime and response time meet SLA thresholds.