Today, all customers who deploy their websites and web apps using Azure Websites generate web logs that track all HTTP transactions and detailed error logging. These logs are stored in Azure Blobs as non-relational files that are semi-structured. Traditionally, semi-structured files are harder to analyze unless they are extracted and transformed into a relational table/database.
Enter Hadoop which was built for these specific use cases like the ability to analyze large volumes of non-relational files. As it relates to web logs, Hadoop can help you understand site metrics like overall web traffic, the length of time someone stays on the site, the effectiveness of certain pages or campaigns, etc.
However, we understand that many organizations have yet to build expertise using Hadoop and need help getting started. For customers to get productive on Hadoop quickly, Azure HDInsight has updated its query console for customers to browse a gallery of templates and be guided through the completion of a task. In this case, the task available is doing analytics on log files captured in Azure Website. We’ll be introducing more tasks in the coming months.
To get started, please visit https://manage.windowsazure.com/
provision an Azure HDInsight Hadoop cluster then select “Query Console” to view the solutions.
From there, you can use Hive to issue a SQL-like query on your web logs. Finally, if you have Excel, you can also visualize your analysis in familiar Excel charts. Learn how you can use Excel with HDInsight here