Optimize querying

With data written to your cluster, you can now begin to define and test your typical query patterns and employ optimizations to ensure query performance.

Define your query patterns

Understanding your typical query pattern helps to prioritize what optimizations can be made to ensure your query performance meets your requirements. For example, consider the following questions:

• Do you typically query data by a specific tag values?
  Apply custom partitioning to your target database or table to partition by those tags. Partitioning by commonly-queried tags helps InfluxDB to quickly identify where the relevant data is in storage and improves query performance.
• Do you query tables with wide schemas?
  Avoid using wildcards (*) in your SELECT statement. Select specific columns you want returned in your query results. The more columns queried, the less performant the query.
• Do you query large, historical time ranges? Use time-based aggregation methods to downsample your data and return aggregate values per interval of time instead of all the data.

Decide on your query language

InfluxDB Clustered supports both SQL and InfluxQL–a SQL-like query language designed for InfluxDB v1 and specifically querying time series data.

SQL

The InfluxDB SQL implementation is a full-featured SQL query engine powered by Apache DataFusion. It benefits from a robust upstream community that is constantly improving the functionality and performance of the engine. Some time series-specific queries (such as time-based aggregates) are more verbose in SQL than in InfluxQL, but they are still possible.

InfluxQL

InfluxQL is designed specifically for time series data and simplifies many time series-related operations like aggregating based on time, technical analysis, and forecasting. It isn’t as full-featured as SQL and requires some understanding of the InfluxDB v1 data model.

Optimize your queries

View the query optimization and troubleshooting documentation for guidance and information on how to troubleshoot and optimize queries that do not perform as expected.

Analyze queries

Both SQL and InfluxQL support the EXPLAIN and EXPLAIN ANALYZE statements that return detailed information about your query’s planning and execution. This can provide insight into possible optimizations you can make for a specific query. For more information, see Analyze a query plan.

Custom-partition data

InfluxDB Clustered lets you define how data is stored to ensure queries are performant. Custom partitioning lets you define how InfluxDB partitions data and can be used to structure your data so it’s easier for InfluxDB to identify where the data you typically query is in storage. For more information, see Manage data partitioning.

Report query performance issues

If you have a query that isn’t meeting your performance requirements despite implementing query optimizations, please following the process described in Report query performance issues to gather information for InfluxData engineers so they can help identify any potential solutions.