Following the recent release of value-based threshold support for SQL Rules, you can now add a sampling query to the monitor to pull a wider set of records at the time of the rule execution to aid in incident investigation.
Monte Carlo now automatically imports Snowflake tags as Monte Carlo tags. These imported tags are regularly updated to match tags in Snowflake, and will appear in the Tags module on the Catalog view for a table. Imported tags can be used for Notification routing.
NOTE: Self-hosted collector customers will need to update their collector version to access this feature.
In addition to defining a Volume SLI by tracking total records in the table, you can now define a Volume SLI by the growth rate of the records added to the table over a given time period.
The Monte Carlo Slack app was reviewed and approved to be added to Slack's directory.
To improve Monte Carlo monitors, we rely heavily on customer feedback to tell us which incidents are most helpful. With this improvement, you can now document feedback and update model thresholds directly from Incident IQ.
Use search operators to refine your search query by specifying table fields, description and tags. Additionally, you can include or exclude specific terms use the + and - prefix.
To provide more context when investigating an incident, we now expose comments from related historical incidents to understand how previous incidents were addressed.
Add quotes to your Catalog search terms to increase precision of search query and support terms with spaces.
Examples:
tags:owner=\"John Doe\"description:\"client information\"\"distribution detectors\"
Released a handful of improvements to Monitors as code to support new features and improve reliability:
- Add support for Freshness/Volume SLIs with multiple tables (example here)
- Use dynamic\_thresholds to enable automatic thresholds in SQL Rules (mentioned in the documentation here)
- Add validation to prevent collisions with generated names when applying a template (see documentation for more detail)
With value based comparisons, you can write a SQL query that returns a single value (ex. sum of revenue over past day) and use the output value direct as the comparison metric.