CTI3561-Lab1.1-William-Feller

Lab Assignment 1.1 – Philosophy of Monitoring This lab will focus on thinking through how we develop a monitoring strategy. Before you begin monitoring you need to think about what you need to know about a system to address potential issues, predict failures, and evaluate trends. This lab will look at questions regarding monitoring design and data analysis. By answering these questions students will prepare to analyze monitors created later in class and answer questions regarding best practices for monitoring. Objective: The student will answer questions related to discussion points from the lecture, readings, and data presented in the lab. The student may speculate as part of the reviewing monitoring data is creating a hypothesis about what is happening then comparing data to the hypothesis to see if the data supports it. Answer each question completely following instructions for each. Task 1: Monitoring Questions Answer the below questions with complete answers. Remember, you can speculate about what you are seeing. Include why you believe your answer is possible. NOTE: Answers should be at least two complete sentences with three to four sentences being more likely for a complete answer. 1. Why is it important to understand all the dependencies that make your system run regarding monitoring? Understanding all dependencies in a system is crucial for effective monitoring because it allows you to identify potential points of failure or performance bottlenecks. By comprehensively monitoring dependencies, you can ensure that critical components are functioning optimally, anticipate issues before they impact the system, and proactively address them to maintain system reliability and performance. 2. Explain a network dependency that should be monitored that if unavailable could impact a web application? 3. An example of a critical network dependency for a web application is the Domain Name System (DNS). If DNS services are unavailable, users won't be able to resolve domain names to IP addresses, effectively rendering the web application inaccessible, as users won't be able to reach the application servers. 4. What types of dependencies may require a service level agreement? A service level agreement is a guarantee that a service will remain available within thresholds. Think about the music streaming service we described in the lecture. What services should have a service level agreement?

Services like authentication, database access, and content delivery for the music streaming application should have SLAs. For instance, the availability and response times of authentication services and content retrieval are critical for a seamless user experience and should be guaranteed within defined thresholds in an SLA. 5. Review the In The News Slide from lecture. What data could have been used to predict the influx of traffic? Can a website be infinitely elastic based on required load? Why or why not? 6. Data like historical website traffic patterns, marketing campaigns, seasonal trends, and events could be used to predict traffic influx. A website's elasticity is not infinite — it's constrained by factors like server capacity, bandwidth, and the application's design. However, cloud-based scaling and efficient load balancing can significantly enhance a website's ability to handle varying traffic loads. 7. Look at the below chart. When do the peaks happen? Speculate on the type of service that may have user’s login that could create this type of trend. What type of service may have this kind of login trend? The peaks in the chart likely indicate login activity. A service that may create such a trend is an e-commerce platform where users log in to make purchases, particularly during peak shopping times like holidays or weekends. 8. Using the chart from question 5 during what months would it be best to perform maintenance on the systems based on the user logins? Which month would be best? Use the data to explain why. It would be best to perform maintenance during the months with lower login activity, which appears to be october and november in the chart. Lower user logins mean less disruption to active users and a reduced risk of impacting critical operations. 0 10 20 30 40 50 60 70 80 90 100 LOGINS (PER THOUSAND USERS)

12. Using the chart from question 9 what could cause the breaks in the trend line? There are a few reasons this may happen. Speculate as to what might cause there to be a gap in data. Try to provide a few reasons. Breaks in the trend line could be due to system reboots, software updates, maintenance activities, or temporary system outages. They could also result from monitoring interruptions or data collection issues, causing gaps in the recorded CPU utilization data. Deliverable: Answer the above questions Lab Write-Up In addition to submitting the answers to the above ten questions you need to answer the below questions with one to two paragraphs minimum . 1. What could you do to validate the metrics you get from a monitoring system are accurate? 2. Can you ever have too much monitoring and data? Why or why not? 1. Validating Accuracy of Monitoring Metrics: To ensure the accuracy of metrics obtained from a monitoring system, you can: • Compare with Ground Truth: Cross-verify the monitoring data with known or expected values. For instance, compare server response times reported by the monitoring system with actual response times measured through controlled tests. • Use Redundant Monitoring: Implement multiple monitoring solutions and cross- reference their data. If different systems report similar metrics, it boosts confidence in the accuracy. 0 10 20 30 40 50 60 70 80 90 100 0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 18:00 19:00 20:00 21:00 22:00 23:00 CPU Utilization - Active Directory

• Perform Regular Audits: Conduct periodic audits of the monitoring system to validate its configuration, data collection mechanisms, and alerting thresholds. • Calibrate Monitoring Devices: Regularly calibrate monitoring devices to ensure they're providing accurate measurements. 2. Having Too Much Monitoring and Data: Yes, it is possible to have too much monitoring and data, and this can lead to problems: • Information Overload: Excessive data can overwhelm teams, making it challenging to focus on critical issues. Sorting through vast amounts of data to find meaningful insights can be time-consuming and inefficient. • Cost and Resource Intensiveness: Storing, processing, and managing extensive monitoring data can be costly in terms of storage infrastructure, processing power, and human resources required to manage and analyze the data. • Alert Fatigue: Too many alerts from excessive monitoring can lead to alert fatigue, where important alerts get ignored or missed amidst the noise of less significant notifications. • Focus on Relevant Metrics: It's important to focus on monitoring metrics that directly impact system performance, availability, and user experience. Prioritize monitoring the critical aspects of your system to avoid drowning in unnecessary data. Striking a balance between comprehensive monitoring and avoiding information overload is crucial to ensure effective monitoring without overwhelming the team or wasting resources. Deliverable: Answer the above questions