Sergey Nivens - Fotolia
Pepperdata Query Spotlight provides insight into data queries
New Query Spotlight tool looks to provide visibility across a data analysis cluster into how a query performs over time and why it might have degraded performance.
Looking beyond just managing data, a key challenge for many organizations is understanding data query performance. Analytics stack performance software vendor Pepperdata is taking aim at the challenge of improving data queries with its new Query Spotlight product.
Multiple vendors and tools are already on the market to help users accelerate big data workflows, including Pepperdata, based in Cupertino, Calif. The vendor's big data performance suite focuses on application profiling to improve data operations.
Diagnosing query performance ills
The new Pepperdata Query Spotlight tool goes a step further, providing users with insight into how database queries are running and identifying potential performance problems as well as structural faults within a database.
Query Spotlight, which became generally available Jan. 28, is a natural step forward for Pepperdata and fills what had been a hole in its lineup, said 451 Research analyst James Curtis.
"Environment and application metrics can get you only so far," Curtis said. "Queries can be nuanced and can at times be a potential runway train, so to speak."
As such, Curtis noted that the Pepperdata Query Spotlight product should be quite useful for enterprises, but he expects its real value is in understanding the context of the query, which can be provided by the Pepperdata's other products.
Pepperdata Query Spotlight features
Pepperdata's core platform is focused on automating performance analysis for big data and analytics stacks, explained Charles Marker, the vendor's vice president of engineering.
James CurtisAnalyst, 451 Research
Pepperdata has a lightweight agent that runs on every node in a data stack and collects more than 300 different metrics every five seconds. The metrics includes information about memory consumption and load average per node. All the metrics are put into a time-series database, which enables Pepperdata to provide correlated metrics about data options.
The Pepperdata platform initially targeted infrastructure, then expanded to provide visibility into applications and is now adding Query Spotlight for analytics queries.
"We will enable you to see what queries are running on the system and to have historical correlated views," Marker said.
An example is a query that runs every day and at some point starts to run slower. With Pepperdata Query Spotlight, the user can look at the query and compare it to past runs, Marker explained.
Query Spotlight also can examine the database EXPLAIN PLAN, which is a database function that determines how a query will be executed. Looking through the EXPLAIN PLAN with Query Spotlight enables users to precisely analyze how a query is running across different nodes in a cluster.
"So with Query Spotlight we can see what all the operators are, what are all the tables and how much data that a query is accessing, to try to understand if maybe there's something that has changed," Marker said.
Marker also noted that users can get a granular view of the database and look at the specific database tables that are being accessed by a given query. With the database view, Pepperdata shows users all the queries that were running against a database table at a given point in time.
"We provide the ability to look at queries and understand from one place what's going on with those queries," Marker said.