Zeek Broker Versions Save

Add new feature that allows users to retrieve a status report about the current state of Broker over HTTP in JSON format. This report gives a more detailed picture than the metrics do, including information about attached stores, the current connection state of a cluster, and more. The endpoint for retrieving this report is /v1/status/json on the same port used for metrics.

Fix a regression where the endpoint::peers() method was always returning a status of connected for all peers. This is most noticeable with zeekctl when collecting and printing peer statuses.

Force static builds with the bundled version of CAF to ensure that all of the dependencies between the various CAF targets are met.

Fix handling of fragmented WebSocket frames

Add interface include dirs to broker lib targets

Minimize messaging in sim_clock::advance_time

Update CAF submodule for SPCP buffer fix

Fix handling of peer/client errors

Implement graceful disconnect handshake

Add synchronous, journal_mode, failure_mode, and integrity_check options to the SQLite store backend.

Fix to use IPv6 mode for accept sockets by default

Add support for using Zeek-style port/protocol notation as arguments

Add support for building with GCC 13

Update macOS CI platforms (Tim Wojtulewicz, Corelight)

Upgrade pybind11 to version 2.10.0 (Dominik Charousset, Corelight)

Fix Windows CI builds (Tim Wojtulewicz, Corelight)

Fix a potential stall in Broker’s internal data pipeline (Dominik Charousset, Corelight)

Fix Windows CI builds (Tim Wojtulewicz, Corelight)

Fix a potential stall in Broker’s internal data pipeline (Dominik Charousset, Corelight)

Fix CMake package file when bundling CAF (Dominik Charousset, Corelight)

Fix a performance bottleneck in Broker-internal caching.

Improve performance with the stealing scheduler policy.

Fix messaging between clones and proxies that was resulting in an error message being reported.

Fix a bug that prevented Broker nodes to recover from OpenSSL errors.

Fix handling of buffer sizes that caused Broker to stall despite having sufficient capacity.

Add missing error signaling

Fix communication between Broker data stores.

Fix a performance bottleneck in Broker-internal caching.

Improve performance with the stealing scheduler policy.

Fix messaging between clones and proxies that was resulting in an error message being reported.