Add Configurable Query Tags for Destination Writes to Improve Resource Usage and Cost Tracking
PlannedRequesting support for automatically applying a configurable query tag to all queries executed by Fivetran in the destination (for example, Snowflake) during syncs, logs, and data ingestion operations.
This enhancement would allow customers to clearly identify Fivetran‑generated activity within destination query history and warehouse metering views. By tagging these queries, teams would be able to:
- Accurately track compute and resource usage attributable to Fivetran syncs
- Improve cost attribution for data ingestion and data movement workloads
- Monitor and analyze warehouse utilization and performance impact caused specifically by Fivetran
- Support FinOps, governance, and capacity planning initiatives
- Simplify auditing and operational visibility without relying on indirect inference methods
An option to define query tags at a global or per‑connector level would provide flexibility for larger environments with multiple connectors and shared warehouses. This feature would significantly enhance observability and cost transparency for customers using Fivetran at scale.
-
Hi Odalis,
Thanks for sharing this request—we appreciate the context. We’re currently working on a feature that will add query tagging for operations generated by Fivetran while loading data into the warehouse.
Could you share a bit more about your need for configurability? Understanding your use case will help us ensure the solution aligns well with your requirements.
Best regards,
-
Hi Egidio,Thanks for the follow-up, happy to clarify our use case.
In practice, the expectation on our side is that Fivetran provides a configurable field to set the query tag, and we will define and manage the tag ourselves.
We would provide the tag value/structure (for example at connector level, potentially including client/source identifiers), and Fivetran would simply apply it to the generated queries.
This is important for us because:- We want to avoid being dependent on a fixed or Fivetran-defined tag format
- If we need to change the tag structure in the future, we should be able to do so without requiring any changes from Fivetran
- It’s also a standard approach in our environment to be able to define query tags as a configurable option across ingestion, processing, reporting, etc.
So the key requirement is flexibility: we should be able to fully control what the query tag looks like, rather than having it generated or constrained by Fivetran.Let me know if you need any more detail on how we would structure the tags on our side. -
Hi Odalis,
Thanks for the detailed context—this is really helpful and gives us a clear picture of your requirements.
At this stage, we’re not planning to support fully configurable or customer-defined query tag formats. Our approach is to provide a consistent tagging structure out of the box, which we expect will cover the majority of use cases.
Introducing configurability both on our side and at the destination often makes ongoing maintenance and updates more complex, so we’re aiming to keep the experience as simple and predictable as possible.
Once this is in place, we’ll evaluate how well it meets customer needs and iterate if there are gaps.
Appreciate you sharing this—it’s very useful input as we move forward.
Best regards,
-
Hi Egidio,
Thanks for the clarification, this makes sense.
For now, a standard query tagging approach works for us, as long as the tags are consistently applied and clearly identifiable, so we can reliably attribute Snowflake credit consumption to Fivetran activity.
-
Hi Fivetran,
One quick follow‑up: Do you have an estimated timeline for when query tags will be implemented?
Also, I’d like to get a better sense of how the tags will be generated once the feature is released. Will the same query tag be applied to every sync, or will each execution receive its own unique tag?
-
Hi Odalis,
Short term, I don’t have an exact date yet as we’re working to ensure the solution is complete and stable.
I'll keep you updated here.
Tags will include sync id, connection, type of sync, and schema.Best,
Please sign in to leave a comment.
Comments
6 comments