Apache Solr 10 has officially been released, with the Solr project announcing Apache Solr 10.0.0 on 3 March 2026. Solr describes itself as a fast, open-source, multi-modal search platform built on Apache Lucene, supporting full-text, vector, and geospatial search. The release is significant because Solr 10 is now the current stable major version, while 9.10.x moves into the role of the previous major line with only occasional critical fixes.
One of the most important technical changes in Solr 10 is the new Java requirement. The official upgrade notes state that Solr 10 requires at least Java 21, while SolrJ 10 requires at least Java 17. This is a major platform shift and one of the first things administrators and developers need to check before planning an upgrade. It means older Java runtimes that may still be common on long-lived servers are no longer enough for the Solr server itself.
Another major theme in Solr 10 is vector search. The official release highlights point to a significant expansion of Solr’s vector capabilities, including support for scalar and binary quantized dense vectors, which are intended to improve performance and reduce memory usage. The release also introduces cuVS-Lucene integration as a pluggable codec for GPU acceleration of vector operations, along with new tuning parameters such as efSearch and efSearchScaleFactor for k-nearest-neighbor workloads. This shows that Solr 10 is not only a classic enterprise search release, but also a serious step toward modern AI and semantic-search use cases.
Observability is another area that changed in a very visible way. Solr 10 has migrated from Dropwizard metrics to OpenTelemetry (OTEL). According to the Solr 10 upgrade notes, this brings native Prometheus support, OTLP export support, tracing correlation through exemplars, and a new model based on attributes and labels. The /admin/metrics API now defaults to Prometheus exposition format instead of the old XML/JSON/javabin output modes, which is an important operational change for monitoring setups.
The server stack itself has also been modernized. Solr 10 upgrades to Jetty 12, replacing older Jetty generations that had reached end-of-life. The project notes that this upgrade brings support for modern HTTP protocols and completes the move to the Jakarta EE 10 namespace, meaning Solr no longer ships with JARs using the old javax namespace. In practice, this is a substantial internal modernization and may matter for integrations, extensions, and deployment environments.
There are also important improvements in stability and infrastructure behavior. The Solr project says that ZooKeeper interactions are now managed through Apache Curator, intended to improve reliability and stability. On the tooling side, the bin/solr CLI has been reworked to behave more consistently, with long options now following Unix-style --long-option formatting. Solr 10 also changes default startup behavior so that bin/solr start now launches in SolrCloud mode by default, while standalone mode requires --user-managed. These are the kinds of changes that may not sound flashy, but they can matter a lot in real production automation and upgrade scripts.
At the same time, Solr 10 removes and deprecates several older components. The release notes say Solr now standardizes on Jetty HttpClient and no longer uses Apache HttpClient internally. It also removes the jaegertracer-configurator, analytics, and hadoop-auth modules, removes the local Tika extraction engine in favor of using TikaServer, and drops deprecated field types such as ExternalFileField, CurrencyField, and EnumField. These removals are important because major version upgrades are exactly where long-deprecated features often disappear.
There are a few other changes worth mentioning. Solr 10 introduces an alternative experimental Admin UI available at /solr/ui, upgrades the official Docker image base to Ubuntu 24, and includes LatestVersionMergePolicyFactory to help maintain index compatibility for future upgrade scenarios. Together, these details show that Solr 10 is not a small maintenance release but a broad modernization across search features, observability, containerization, and operational tooling.
In summary, Solr 10 is a major release aimed at modern workloads and modern infrastructure. It raises the Java baseline, deepens vector-search and GPU-related capabilities, replaces older observability with OpenTelemetry, moves to Jetty 12 and Jakarta, and removes a number of legacy components. For organizations already using Solr, this release looks powerful, but it is clearly one that should be approached with proper upgrade planning rather than treated like a routine point update.
https://solr.apache.org/
https://solr.apache.org/guide/solr/late ... lr-10.html