Download the release on GitHub.
Features
Section titled “Features”Allow assigning ids for pipelines manually
Section titled “Allow assigning ids for pipelines manually”The /pipeline/create and /pipeline/launch endpoints now accept an optional
id parameter for assigning the pipeline’s ID manually.
By @dominiklohmann in #5026.
Bug Fixes
Section titled “Bug Fixes”Make drop not remove empty records
Section titled “Make drop not remove empty records”Dropping all fields from a record with the drop operator no longer removes the
record itself. For example, from {x: {y: 0}} | drop x.y now returns {x: {}}
instead of {}.
By @dominiklohmann in #5021.
Fix crash in the MSB in merging mode
Section titled “Fix crash in the MSB in merging mode”We fixed a bug in the read_xsv and parse_xsv family of operators and
functions that caused the parser to fail unexpectedly when the data contained
a list (as specified through the list separator) for fields where the provided
schema did not expect lists.
By @IyeOnline in #5028.
Fix tls option of to_splunk
Section titled “Fix tls option of to_splunk”Using the tls option of the to_splunk operator no longer directly emits an
error.
Make UDO resolution order-independent
Section titled “Make UDO resolution order-independent”The resolution of user-defined operator aliases in the tenzir.operators
section is no longer order-dependent. Previously, an operator foo may have
depended on an operator bar, but not the other way around. This limitation no
longer exists.
By @dominiklohmann in #5029.
Fix startup delay in from_fluent_bit
Section titled “Fix startup delay in from_fluent_bit”We fixed a bug that caused pipelines with from_fluent_bit to not report their
startup successfully, causing errors when deploying pipelines starting with the
operator through the Tenzir Platform.
By @dominiklohmann in #5025.