Conversation
abdonpijpelink
added
do not merge
✅ Deploy Preview for condescending-goldwasser-91acf0 ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
mrscoopers
requested changes
Feb 16, 2026
|
|
||
| Writing queries is hard: users often struggle to precisely formulate search queries. At the same time, judging the relevance of a given search result is often much easier. Retrieval systems can leverage this [relevance feedback](/articles/search-feedback-loop/) to iteratively refine results toward user intent. | ||
|
|
||
| This release introduces a new [Relevance Feedback Query](/documentation/concepts/search-relevance/#relevance-feedback) as a scalable, vector‑native approach to incorporating relevance feedback. The Relevance Feedback Query uses a small amount of model‑generated feedback to guide the retriever through the entire vector space, effectively nudging search toward “more relevant” results without requiring expensive loops, large models, or human labeling. This enables the engine to traverse billions of vectors with improved recall without having to retrain models. |
Contributor
There was a problem hiding this comment.
large models
(they can use them for feedback, but benefit here is that these models won't have to rerank/score/judge hundreds of examples)
Maybe expensive retrievers?
mrscoopers
reviewed
Feb 17, 2026
|
|
||
| This release introduces a new [Relevance Feedback Query](/documentation/concepts/search-relevance/#relevance-feedback) as a scalable, vector‑native approach to incorporating relevance feedback. The Relevance Feedback Query uses a small amount of model‑generated feedback to guide the retriever through the entire vector space, effectively nudging search toward “more relevant” results without requiring expensive loops, expensive retrievers, or human labeling. This enables the engine to traverse billions of vectors with improved recall without having to retrain models. | ||
|
|
||
| This method works by collecting lightweight feedback on just a few top results, creating “context pairs” of more‑ and less‑relevant examples. These pairs define a signal that adjusts the scoring function during the next retrieval pass. Instead of rewriting queries or rescoring large batches of documents, Qdrant modifies how similarity is computed. [Experiments](/articles/relevance-feedback) demonstrate substantial gains, especially when pairing expressive retrievers with strong feedback models. |
Contributor
There was a problem hiding this comment.
Lat nitpick: maybe we could reference the article more explicitly, aka
- In the article "XYZ" [link], you can understand the idea behind the method & experiments
- In the docs [link], you can get the instructions on using the method right away
Wdyt?
timvisee
reviewed
Feb 19, 2026
Member
timvisee
left a comment
timvisee
left a comment
There was a problem hiding this comment.
Some remarks. Take them with a grain of salt.
Comment on lines
21
to
25
| **Relevance Feedback Query:** Improve the quality of search results by incorporating information about their relevance. | ||
|
|
||
| **Search Latency Improvements:** Two search latency improvements: a change to improve search latency under high write loads, and delayed fan-outs to reduce tail latency. | ||
|
|
||
| **Greater Operational Observability:** Improved insights into operational metrics and faster troubleshooting with a new cluster-wide telemetry API and segment optimization monitoring. |
Member
There was a problem hiding this comment.
Imo the word "improve" is used a lot in the first few sections.
Also I don't like this sum up to much
Two search latency improvements:
as I believe there is more to it.
Suggested change
| **Relevance Feedback Query:** Improve the quality of search results by incorporating information about their relevance. | |
| **Search Latency Improvements:** Two search latency improvements: a change to improve search latency under high write loads, and delayed fan-outs to reduce tail latency. | |
| **Greater Operational Observability:** Improved insights into operational metrics and faster troubleshooting with a new cluster-wide telemetry API and segment optimization monitoring. | |
| **Relevance Feedback Query:** Tune the quality of search results by incorporating information about their relevance. | |
| **Search Latency Improvements:** Manage your search latency with new tools such as the update queue and delayed fan-outs to reduce tail latency. And benefit from many internal search performance improvements. | |
| **Greater Operational Observability:** Better insights into operational metrics and faster troubleshooting with a new cluster-wide telemetry API and segment optimization monitoring. |
The above is definitely not final. But maybe some of it is useful. Thoughts?
Member
Author
There was a problem hiding this comment.
Great feedback. I've incorporated your suggestions, but kept the first "improve" in "Improve the quality of search results...".
Comment on lines
89
to
94
| <figure> | ||
| <img width="75%" src="/blog/qdrant-1.17.x/optimizer-web-ui.png"> | ||
| <figcaption> | ||
| The new user interface in the Web UI provides an overview of the current optimization status and a timeline of current and past optimization cycles. | ||
| </figcaption> | ||
| </figure> |
Member
There was a problem hiding this comment.
This view is cluster-wide now, and not just for a single node. I think it's worth to mention.
Comment on lines
135
to
136
| In Qdrant Cloud, navigate to the Cluster Details screen and select Version 1.17 from the dropdown menu. The upgrade process may take a few moments. | ||
| We recommend upgrading versions one by one, for example, 1.15->1.16->1.17. On Qdrant Cloud, the required intermediate updates are automatically performed to ensure a supported upgrade path. |
Member
There was a problem hiding this comment.
I'd swap this around and say something like:
on Qdrant Cloud your upgrade is handled automatically once you hit the button. If you've deployed Qdrant yourself, make sure to hit the latest patch for each minor version. E.g.: 1.15.5->1.16.3->1.17.0
Also notice I mention the latest patch version here.
List of versions can be found here: https://github.com/qdrant/qdrant/tags
|
|
||
| This release includes several changes that reduce search latency. To improve query response times in environments with high write loads, Qdrant can now be configured to avoid creating large unoptimized segments. Additionally, delayed fan-outs help reduce tail latency by querying a second replica if the first does not respond within a configurable latency threshold. | ||
|
|
||
| ### Improved Search Performance Under High Write Loads |
Member
There was a problem hiding this comment.
I find the title quite long. Not sure about this suggestion:
Suggested change
| ### Improved Search Performance Under High Write Loads | |
| ### Make write loads manageable |
Member
Author
There was a problem hiding this comment.
How about simply "Search Latency Under Write Load" for a shorter title?
|
|
||
| ### Improved Search Performance Under High Write Loads | ||
|
|
||
| A common pattern with search engines like Qdrant involves periodically refreshing data from an external source of truth, for example, using nightly batch updates. Newly ingested data needs to be indexed, which is a resource-intensive operation. When the data ingestion rate exceeds the indexing rate, this can lead to issues such as: |
Member
There was a problem hiding this comment.
Suggested change
| A common pattern with search engines like Qdrant involves periodically refreshing data from an external source of truth, for example, using nightly batch updates. Newly ingested data needs to be indexed, which is a resource-intensive operation. When the data ingestion rate exceeds the indexing rate, this can lead to issues such as: | |
| A common pattern with vector search engines like Qdrant involves bulk uploads. Periodically refreshing data from an external source of truth, for example, using nightly batch updates. Newly ingested data needs to be indexed, which is a resource-intensive operation. When the data ingestion rate exceeds the indexing rate, this can lead to issues such as: |
mrscoopers
approved these changes
Feb 19, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Preview