This alert is triggered when the heap memory utilisation of an Amazon Kinesis Data Analytics application exceeds 80%. High heap memory utilisation can lead to potential performance issues, memory leaks, or application crashes if not addressed promptly. This alert indicates that the application is using a significant portion of its allocated memory, and action may be required to prevent adverse effects on performance and stability This alert triggers when heap Memory Utilization exceeds exceeds 80%. Customization Guidance: - Threshold: The default threshold is set at at 80%. Adjust this depending on the specific requirements and behaviour of your application. - Monitoring Period: The default monitoring period is set to 10 minutes but can be adjusted to shorter or longer intervals based on traffic patterns and the criticality of data processing. Configure the monitoring period to an appropriate duration to ensure detection of memory utilisation issues without causing excessive alerting. - Notification Frequency: Balance the alert frequency to optimise responsiveness while minimising noise. Adjust according to the criticality of uninterrupted operation. Action: If this alert is triggered, investigate the cause of the high memory utilisation. This may involve optimising the application code, scaling the application resources, and/or reviewing the application configuration to ensure efficient memory usage.

This alert is triggered when the total memory utilisation of an Amazon Kinesis Data Analytics application exceeds 80%. High memory utilisation indicates that the application or the underlying system is consuming a large portion of the available memory, which can lead to degraded performance, system instability, and potential crashes if not addressed promptly. This alert encompasses all types of memory usage, including heap memory, non-heap memory, and memory used by other processes on the instance. Customization Guidance: - Threshold: The default threshold is set at at 80%. Adjust this depending on the specific requirements and behaviour of your application. - Monitoring Period: The default monitoring period is set to 10 minutes but can be adjusted to shorter or longer intervals based on traffic patterns and the criticality of data processing. Configure the monitoring period to an appropriate duration to ensure detection of memory utilisation issues without causing excessive alerting. - Notification Frequency: Balance the alert frequency to optimise responsiveness while minimising noise. Adjust according to the criticality of uninterrupted operation. Action: If this alert is triggered, investigate the cause of the high memory utilisation. This may involve optimising the application code, scaling the application resources, and/or reviewing the application and/or system configuration to ensure efficient memory usage.

This alert is triggered when the CPU utilisation of an Amazon Kinesis Data Analytics application exceeds 80%. High CPU utilisation indicates that the application is consuming a large portion of the available processing power, which can lead to degraded performance, increased latency, and potential timeouts or failures if not addressed promptly. This alert is crucial for maintaining the performance and reliability of the application, as consistently high CPU usage can signal issues such as inefficient processing, excessive data load, or suboptimal resource allocation. Customization Guidance: - Threshold: The default threshold is set to trigger when CPU Utilisation exceeds exceeds 80%. Set the threshold value to detect when the CPU utilisation exceeds acceptable limits. - Monitoring Period: The monitoring period is set to 10 minutes but can be adjusted to shorter or longer intervals based on traffic patterns and the criticality of data processing. Configure the monitoring period to an appropriate duration to ensure detection of CPU utilisation issues without causing excessive alerting. - Notification Frequency: Adjust the frequency of this alert to balance responsiveness and alert fatigue. Tune it according to the criticality of continuous, uninterrupted data processing for your service. Action: f this alert is triggered, investigate the cause of the high memory utilisation. This may involve optimising the application code, scaling the application resources, or reviewing the configuration to ensure efficient memory usage.

This alert indicates that the age of the oldest record in the Amazon Kinesis Data Analytics application has exceeded the threshold of 1000 milliseconds (1 second). The “oldest record age” represents the time duration since the oldest record in the stream was added. This could be due to multiple reasons including the data processing application experiencing lag, insufficient processing resources, bottlenecks in the data processing pipeline and unusually high volume of incoming data. Delays in processing can result in stale data being delivered to downstream systems, negatively impacting real-time analytics and/or system overload or application failure. Customization Guidance: - Threshold: The default threshold is set to trigger when the average lag time exceeds exceeds 1000ms (1 second). Set the threshold value to detect when the age of the oldest record exceeds acceptable limits. - Monitoring Period: The monitoring period is set to 10 minutes but can be adjusted to shorter or longer intervals based on traffic patterns and the criticality of data processing. - Notification Frequency: Adjust the frequency of this alert to balance responsiveness and alert fatigue. Tune it according to the criticality of continuous, uninterrupted data processing for your service. Action: If this alert is triggered, investigate the cause of the increased lag. This may involve increasing the processing power or parallelism of the Kinesis Data Analytics application to handle the load more effectively, reviewing and optimising the processing logic to reduce bottlenecks and improve efficiency and keeping an eye on the data input rate and checking if any data sources are sending unusually large or complex records that could be slowing down processing.