This release is recommended for all users. Significant performance improvements to query performance and start up time are included as well as improved analysis of PLCrashReporter errors. New features include per-project storage sampling configuration, metrics import, multiple aggregation filters, HTTP/2 support and more.
- Improvements to the default callstack extracted from PLCrashReporter files to include last exception callstacks.
- Support for multiple aggregation filters has been added.
- Support for object submission over the console listener is now supported.
- All incoming network traffic is managed by nginx out of the box, allowing users to leverage nginx features for controlling incoming network traffic.
- Queries that look for a specific fingerprint that has been merged in a group now resolve to that group.
- Performance has been improved for the web console by reducing latency to commonly accessed configuration APIs.
- Improves module detection in the presence of duplicate malformed module entries in a minidump file.
- Significantly improved query performance time. Queries that involve millions of events are now up to four times as fast. Repeated queries for similar aggregates and filtering criteria (e.g., small changes to the time aperture) can reuse internal aggregation and grouping results.
- Default database settings will parallelize database queries more aggressively based on core count.
- Significantly improved start-up time for databases containing many large strings.
- hwloc is no longer required.
- Various minor bug fixes and improvements to stability.