Failing Queries: Queries that are syntactically correct and thus run but do not complete due to lack of compute or time resources. Similar to Zero Impact, Failing Queries incur cost without benefit. Examples include queries that timeout due to disk spillage, that are blocked or are terminated by governors due to run time.
Poorly performing queries are the bane of every data team requiring so much time, manual effort, and expertise. Data engineering teams spend significant efforts on writing batch jobs and manually optimizing their queries and database schema.
Inefficient or poorly written queries are the biggest factors affecting Snowflake’s compute costs. A common approach taken by most organizations is to have a small team of highly experienced DBAs constantly monitor the query logs. They identify the lowest performing queries and then intervene to optimize them.
Continuous efficiency is the gold standard for teams who want to balance costs without slowing down innovation. To achieve this, you will need a way to continuously monitor and remediate data cloud workloads. Many organizations today take the approach of doing point-in-time optimizations, meaning they have a steady cadence at which they review workloads costs and do something about them. Most often, this is monthly or quarterly. We call this “discrete” efficiency rather than continuous, to use the mathematical term. Discrete means “point-in-time” whereas continuous means “all the time”.
To achieve truly continuous efficiency, organizations need a modern AI-powered platform because humans simply can’t keep up with the sheer amount of information that needs to be processed to make efficiency decisions. It’s the reason that so many organizations today only undertake these efficiency exercises every few months.