AWS Certified Data Analytics - Specialty

Using Amazon Kinesis Data Firehose to stream data to Amazon Redshift is the best option for a financial services company that needs to aggregate daily stock trade data and occasionally modify data using SQL. Kinesis Data Firehose reliably loads streaming data into Redshift with minimal latency, and Amazon Redshift supports complex, analytical queries with low latency which is crucial for business intelligence use cases. Additionally, using Amazon QuickSight with Amazon Redshift as the data source enables the creation of business intelligence dashboards to visualize anomalies in stock prices effectively.

To secure access from an on-premises Active Directory to Amazon QuickSight, using an Active Directory connector and single sign-on (SSO) in a corporate network environment is the correct approach. This method leverages the existing Active Directory infrastructure for authentication, allowing users to log in with their corporate credentials. Additionally, it enables IT administrators to manage access to Amazon QuickSight through their existing Active Directory, thereby aligning with the company's security policies and ensuring a seamless and secure user experience.

To make HBase data highly available, it is essential to ensure both data availability and fault tolerance. The most effective approach is to store the data on an EMR File System (EMRFS) instead of HDFS, as EMRFS uses Amazon S3 which is highly durable. Configuring a primary EMR HBase cluster with multiple master nodes ensures that the cluster can withstand master node failures. Additionally, creating a secondary EMR HBase read-replica cluster in a separate Availability Zone provides redundancy and fault tolerance against Availability Zone failures. This setup allows both clusters to access the same HBase root directory in the S3 bucket, ensuring data availability and lowering costs by avoiding redundant data storage. This architecture effectively balances cost and high availability requirements.

To meet the requirement of collecting and analyzing logs with minimal delays, a streaming solution integrated with a search and visualization tool is necessary. Using the Amazon Kinesis Producer Library (KPL) agent on Amazon EC2 to collect and send data to Kinesis Data Firehose, the data can then be pushed to Amazon OpenSearch Service (formerly Amazon Elasticsearch Service) and visualized using OpenSearch Dashboards (Kibana). This option ensures that logs are collected in near real-time, processed efficiently, and made available for visualization with minimal delays. Although minor discrepancies exist in terminology, the described method closely aligns with the desired technical outcome and highly available infrastructure.

To improve the execution time of an AWS Glue job without overprovisioning, enable job metrics to estimate the number of data processing units (DPUs) needed. Based on these metrics, increase the value of the maximum capacity job parameter. This approach helps allocate the precise amount of resources required for the job, improving performance while avoiding unnecessary resource usage. Enabling job bookmarks is more about maintaining state information and preventing the reprocessing of old data, which does not directly impact performance. Adjusting parameters like executor cores or memory overhead depends on specific error codes, which are not present in this scenario.