[In preview] Public Preview: Evaluations with Intelligent Trace Sampling

Vendor: Microsoft
Product: Microsoft Foundry
Type: Pricing
Announcement date: 2026-06-03
Effective date: Date not published
Impact: Affected

In brief

What the source says

Microsoft Foundry observability introduces intelligent trace filtering and sampling for evaluations in public preview. Instead of running evaluations against all production traces, Foundry now select a representative subset of traces using a multi-stage MinHash farthest-first diversity algorithm, including deduplication, hard filers, aggregation and diversity-based selection for evaluation. This approach improves cost efficiency while producing significantly higher lexical diversity and diverse coverage compared to random sampling. Intelligent sampling is particularly effective for: evaluation and benchmarks, rubric generation and finetuning dataset curation. Learn more .

Continue from here

See other Microsoft Foundry changes See the public price for Microsoft Foundry Compare with my estate Ask a question

Read the primary source

Choose your reading

The role changes the reading angle, not the facts, date or level of evidence.

Reading for a salesperson

Who is affected: The audience described by the Microsoft update is represented by: [In preview] Public Preview: Evaluations with Intelligent Trace Sampling Microsoft Foundry observability introduces intelligent trace filtering and sampling for evaluations in public preview. Instead of running evaluations against all production traces, Foundry now select a representative subset of traces using a multi-stage MinHash farthest-first diversity algorithm, including deduplication, hard filers, aggregation and diversity-based selection for evaluation. This approach improves cost effic

Why it matters: The change matters because the official source describes: [In preview] Public Preview: Evaluations with Intelligent Trace Sampling Microsoft Foundry observability introduces intelligent trace filtering and sampling for evaluations in public preview. Instead of running evaluations against all production traces, Foundry now select a representative subset of traces using a multi-stage MinHash farthest-first diversity algorithm, including deduplication, hard filers, aggregation and diversity-based selection for evaluation. This approach improves cost effic

Urgency: Monitor

Next action: Review affected accounts with the customer using the official announcement.

Commercial opportunities

Use the verified change to open a scoped customer conversation.

Technical actions

Assess whether the documented change intersects the customer's current stack.

Questions to ask the customer

Does this documented change affect a product or workload in scope?

Risks and objections

The official source does not establish facts beyond the quoted material.

Points to confirm

The effective date is unknown and must be confirmed before scheduling action.
The affected customer scope is unknown and must be confirmed before action.

Evidence and traceability

Each excerpt is linked to the primary source and retained for verification.

Raw capture
2026-08-30T06:15:07.776356+00:00
Event
[In preview] Public Preview: Evaluations with Intelligent Trace Sampling Microsoft Foundry observability introduces intelligent trace filtering and sampling for evaluations in public preview. Instead of running evaluations against all production traces, Foundry now select a representative subset of traces using a multi-stage MinHash farthest-first diversity algorithm, including deduplication, hard filers, aggregation and diversity-based selection for evaluation. This approach improves cost efficiency while producing significantly higher lexical diversity and diverse coverage compared to random sampling. Intelligent sampling is particularly effective for: evaluation and benchmarks, rubric generation and finetuning dataset curation. Learn more .
source
Product Microsoft Foundry
[In preview] Public Preview: Evaluations with Intelligent Trace Sampling Microsoft Foundry observability introduces intelligent trace filtering and sampling for evaluations in public preview. Instead of running evaluations against all production traces, Foundry now select a representative subset of traces using a multi-stage MinHash farthest-first diversity algorithm, including deduplication, hard filers, aggregation and diversity-based selection for evaluation. This approach improves cost efficiency while producing significantly higher lexical diversity and diverse coverage compared to random sampling. Intelligent sampling is particularly effective for: evaluation and benchmarks, rubric generation and finetuning dataset curation. Learn more .
source
Insight a salesperson
[In preview] Public Preview: Evaluations with Intelligent Trace Sampling Microsoft Foundry observability introduces intelligent trace filtering and sampling for evaluations in public preview. Instead of running evaluations against all production traces, Foundry now select a representative subset of traces using a multi-stage MinHash farthest-first diversity algorithm, including deduplication, hard filers, aggregation and diversity-based selection for evaluation. This approach improves cost efficiency while producing significantly higher lexical diversity and diverse coverage compared to random sampling. Intelligent sampling is particularly effective for: evaluation and benchmarks, rubric generation and finetuning dataset curation. Learn more .
source

Back to the public feed