Identifying suspect queries
Daily, thousands of queries execute on a Teradata database. Some queries' execution is under milliseconds; these are quick and not visible in the monitoring portlet called viewpoint. Some queries have a tendency to execute over minutes, some execute over hours. Whatever the queries are, they need to consume resources on the system. Queries that consume resources such as CPU and IO in their threshold values are classified as good or optimal queries.
Queries that over consume the resources are called suspected queries. Analysis of these queries is necessary from time to time. If these are not inspected over time it can cause the system to under perform.
Definition of and classification of these suspect queries depends on the system. A Teradata database with high configurations such as a large number of nodes can withstand high resource queries for some time, but still these queries—irrespective of the system—are bad in nature and need to captured and tuned.
