2.1

Changes since v2.0:

• Coroutines stack usage is now shown in the engine.coro-pool.stack-usage.max-usage-percent metric. Improved stack-overflow diagnostics.
• HTTP server and HTTP client now support ZSTD decompression. Thanks to Илья Оплачкин and VScdr for the PR!
• Added redis::MakeBulkHedgedRedisRequestAsync() and redis::MakeBulkHedgedRedisRequest().
• OpenTelemetry parent span-id is now passed through AMQP headers along with trace-id. Thanks to TertiumOrganum1 for the PR!
• ugrpc::server::MiddlewareBase now has CallRequestHook and CallResponseHook for intercepting requests and responses.
• components::LoggableComponentBase was renamed to components::ComponentBase. components::RawComponentBase was published.
• Multiple improvements for logging in testsuite.
• gRPC metrics are now not written for methods that were not used at runtime.
• Mongo pools now can be adjusted at runtime via dynamic config MONGO_CONNECTION_POOL_SETTINGS. Congestion Control for individual Mongo databases now could be controlled via MONGO_CONGESTION_CONTROL_DATABASES_SETTINGS. Congestion Control is now enabled by default.
• Reduced contention in coro::Pool and added some tests and benchmarks. Many thanks to Egor Bocharov for the PRs!
• Added urabbitmq::ConsumerComponentBase::Process() accepting the whole urabbitmq::ConsumedMessage. Thanks to TertiumOrganum1 for the PR!
• human_logs.py now supports more options and has more examples and docs embedded. Thanks to TertiumOrganum1 for the PR!
• server::http::HttpStatus and client::http::Status are now aliases to http::StatusCode. Many thanks to SidorovichPavel for the PR!
• Docs and build:
  • find_package(userver) now implicitly calls userver_setup_environment(), includes all the helper CMake userver scripts, making the configuration simpler. Added diagnostics and fix-it hints for some of the CMake missuses.
  • In docs Ctrl+k hotkey now focuses on Search input. Many thanks to Fedor Alekseev for the PR!
  • ODR-violations are now avoided if the userver is built with different standard version than the service.
  • Each sample is now usable as a root project.
  • Each driver now has a @ref QUALITY_TIERS "Quality Tier".
  • Fixed minimal version requirements for Pythons gRPC modules. Thanks to
    Nikita for the PR!
  • Reduced build times by avoiding inclusion of heavy headers.
  • Multiple improvements for docs, build and CI scripts.

2.0

Big new features since the v1.0.0:

• Simplified dynamic configs and embedded defaults into the code.
• Added PostgreSQL connection pools auto-configuration.
• Added YDB driver and basic Kafka driver.
• LISTEN/NOTIFY support for PostgreSQL
• New landing page for the website
• Significantly reduced network data transmission for PostgreSQL
• Supported install in CMake and CPack packaging.
• Implemented middlewares for HTTP server, most of the HTTP server functionality
  was moved to middlewares.
• Improved documentation, added more samples and descriptions.
• Numerous optimizations and build improvements.

Ready to use Ubuntu 22.04 x64 package in attachment.

1.0.0

Big new features since the Beta announcement:

• Implemented WebSockets server
• Added MySQL driver
• RabbitMQ drived was added
• Implemented TLS server
• Enabled PostgreSQL pipelining
• Implemented and enabled Deadline Propagation
• Improved experience with metrics. Added Prometheus and Graphite metric
  formats. Provided a modern simple interface to write and test metrics.
• Added chaos tests for drivers
• Implemented streaming API for the HTTP
• Improved documentation, added more samples and descriptions, improved search.
• Numerous optimizations.
• Numerous build improvements, including Conan and Docker support.

Optimized and improved features that were available at the Beta announcement:

• gRPC client and server
• Mongo driver
• Redis driver
• PostgreSQL
• HTTP server and client
• Logging and Tracing
• ... and many other features.