File python-opentelemetry-exporter-otlp-proto-http.changes of Package python-opentelemetry-exporter-otlp-proto-http
-------------------------------------------------------------------
Sun Sep 21 15:02:05 UTC 2025 - Dirk Müller <dmueller@suse.com>
- update to 1.37.0:
* Add experimental composite samplers
* Add new environment variables to the SDK `OTEL_PYTHON_EXPORTE
R_OTLP_{HTTP/GRPC}_{METRICS/TRACES/LOGS}_CREDENTIAL_PROVIDER`
that can be used to inject a `requests.Session` or
`grpc.ChannelCredentials` object into OTLP exporters created
during auto instrumentation #4689.
* Filter duplicate logs out of some internal `logger`'s logs on
the export logs path that might otherwise endlessly log or
cause a recursion depth exceeded issue in cases where logging
itself results in an exception. (#4695).
* docs: linked the examples with their github source code
location and added Prometheus example
* Permit to override default HTTP OTLP exporters headers
* semantic-conventions: Bump to 1.37.0
* opentelemetry-sdk: fix handling of OTEL_ATTRIBUTE_COUNT_LIMIT
in logs
* Performance: Cache `importlib_metadata.entry_points`
* opentelemetry-sdk: fix calling Logger.emit with an API
LogRecord instance
- update to 1.36.0:
* Add missing Prometheus exporter documentation
* Overwrite logging.config.fileConfig and
logging.config.dictConfig to ensure the OTLP `LogHandler`
remains attached to the root logger. Fix a bug that can cause
a deadlock to occur over `logging._lock` in some cases
(#4636).
* otlp-http-exporter: set default value for param `timeout_sec`
in `_export` method
* Update OTLP gRPC/HTTP exporters: calling shutdown will now
interrupt exporters that are sleeping before a retry attempt,
and cause them to return failure immediately. Update
BatchSpan/LogRecordProcessors: shutdown will now complete
after 30 seconds of trying to finish exporting any buffered
telemetry, instead of continuing to export until all
telemetry was exported. (#4638).
- update to 1.35.0:
* Update OTLP proto to v1.7 #4645.
* Add `event_name` as a top level field in the `LogRecord`.
Events are now simply logs with the `event_name` field set,
the logs SDK should be used to emit events (#4652).
* Update OTLP gRPC/HTTP exporters: the export timeout is now
inclusive of all retries and backoffs. A +/-20% jitter was
added to all backoffs. A pointless 32 second sleep that
occurred after all retries had completed/failed was removed.
(#4564).
* Update ConsoleLogExporter.export to handle LogRecord's
containing bytes type in the body (#4614).
* opentelemetry-sdk: Fix invalid `type: ignore` that causes
mypy to ignore the whole file
* Add `span_exporter` property back to `BatchSpanProcessor`
class
* Fix license field in pyproject.toml files
* Update logger level to NOTSET in logs example
* Logging API accepts optional `context`; deprecates
`trace_id`, `span_id`, `trace_flags`. (#4597) and
* sdk: use context instead of trace_id,span_id for initializing
LogRecord
* Rename LogRecordProcessor.emit to on_emit
* Logging API hide std_to_otel function to convert python
logging severity to otel severity
* proto: relax protobuf version requirement to support v6
* Bump semantic-conventions to 1.36.0
* Set expected User-Agent in HTTP headers for grpc OTLP
exporter
- update to 1.34.0:
* typecheck: add sdk/resources and drop mypy
* Refactor `BatchLogRecordProcessor` to simplify code and make
the control flow more clear (#4562 and #4535).
* Use PEP702 for marking deprecations
* Refactor `BatchLogRecordProcessor` and `BatchSpanProcessor`
to simplify code and make the control flow more clear (#4562
#4535, and #4580).
* Remove log messages from `BatchLogRecordProcessor.emit`, this
caused the program to crash at shutdown with a max recursion
error (#4586).
* Configurable max retry timeout for grpc exporter
* opentelemetry-api: allow importlib-metadata 8.7.0
* opentelemetry-test-utils: assert explicit bucket boundaries
in histogram metrics
* Bump semantic conventions to 1.34.0
* Drop support for Python 3.8
-------------------------------------------------------------------
Mon May 26 07:21:12 UTC 2025 - Nico Krapp <nico.krapp@suse.com>
- Update to 1.33.1
* Remove log messages from BatchLogRecordProcessor.emit, this caused the
program to crash at shutdown with a max recursion error
- Update to 1.33.0
* Fix intermittent Connection aborted error when using otlp/http exporters
* opentelemetry-sdk: use stable code attributes:
code.function -> code.function.name, code.lineno -> code.line.number,
code.filepath -> code.file.path
* Fix serialization of extended attributes for logs signal
* Handle OTEL_PROPAGATORS contains None
* docs: updated and added to the metrics and log examples
* Bump semantic conventions to 1.33.0
-------------------------------------------------------------------
Tue Feb 25 22:23:42 UTC 2025 - Dirk Müller <dmueller@suse.com>
- update to 1.30.0:
* Always setup logs sdk,
OTEL_PYTHON_LOGGING_AUTO_INSTRUMENTATION_ENABLED only
controls python `logging` module handler setup
* Add `attributes` field in `metrics.get_meter` wrapper
function
* Add Python 3.13 support
* sdk: don't log or print warnings when the SDK has been
disabled
* Fix span context manager typing by using ParamSpec from
typing_extensions
* Fix serialization of None values in logs body to match
1.31.0+ data model
* [BREAKING] semantic-conventions: Remove `opentelemetry.semcon
v.attributes.network_attributes.NETWORK_INTERFACE_NAME`
introduced by mistake in the wrong module.
* Add support for explicit bucket boundaries advisory for
Histograms
* semantic-conventions: Bump to 1.30.0
-------------------------------------------------------------------
Wed Jan 8 10:44:43 UTC 2025 - John Paul Adrian Glaubitz <adrian.glaubitz@suse.com>
- Update to 1.29.0
* Fix crash exporting a log record with None body
* Fix metrics export with exemplar and no context and filtering observable instruments
* Fix recursion error with sdk disabled and handler added to root logger
* sdk: setup EventLogger when OTEL_PYTHON_LOGGING_AUTO_INSTRUMENTATION_ENABLED is set
* api: fix logging of duplicate EventLogger setup warning
* sdk: fix setting of process owner in ProcessResourceDetector
* sdk: fix serialization of logs severity_number field to int
* Remove `TestBase.assertEqualSpanInstrumentationInfo` method,
use `assertEqualSpanInstrumentationScope` instead
* sdk: instantiate lazily `ExemplarBucket`s in `ExemplarReservoir`s
* semantic-conventions: Bump to 1.29.0
- from version 1.28.0
* Removed superfluous py.typed markers and added them where they were missing
* Include metric info in encoding exceptions
* sdk: Add support for log formatting
* sdk: Add Host resource detector
* sdk: Implementation of exemplars
* Implement events sdk
* Update semantic conventions to version 1.28.0
* Add support to protobuf 5+ and drop support to protobuf 3 and 4
* Update environment variable descriptions to match signal
* Record logger name as the instrumentation scope name
* Fix memory leak in exporter and reader
* Drop `OTEL_PYTHON_EXPERIMENTAL_DISABLE_PROMETHEUS_UNIT_NORMALIZATION` environment variable
* Improve compatibility with other logging libraries that override
`LogRecord.getMessage()` in order to customize message formatting
-------------------------------------------------------------------
Wed Sep 11 08:23:52 UTC 2024 - Dirk Müller <dmueller@suse.com>
- update to 1.27.0:
* Implementation of Events API
* Make log sdk add `exception.message` to logRecord for
exceptions whose argument is an exception not a string
message
* Fix use of `link.attributes.dropped`, which may not exist
* Running mypy on SDK resources
* Added py.typed file to top-level module
* Drop Final annotation from Enum in semantic conventions
* Update log export example to not use root logger
* sdk: Add OS resource detector
* sdk: Accept non URL-encoded headers in
`OTEL_EXPORTER_OTLP_*HEADERS` to match other languages SDKs
* Update semantic conventions to version 1.27.0
* Add support to type bytes for OTLP AnyValue
* Export ExponentialHistogram and ExponentialHistogramDataPoint
* Implement Client Key and Certificate File Support for All
OTLP Exporters
* Remove `_start_time_unix_nano` attribute from
`_ViewInstrumentMatch` in favor of using `time_ns()` at the
moment when the aggregation object is created
- update to 1.26.0:
* Standardizing timeout calculation in measurement consumer
collect to nanoseconds
* optional scope attributes for logger creation
* optional scope attribute for tracer creation
* OTLP exporter is encoding invalid span/trace IDs in the logs
fix
* Update sdk process resource detector `process.command_args`
attribute to also include the executable itself
* Fix `start_time_unix_nano` for delta collection for explicit
bucket histogram aggregation
* Fix `start_time_unix_nano` for delta collection for sum
aggregation
* Update opentracing and opencesus docs examples to not use
JaegerExporter
* Do not execute Flask Tests in debug mode
* When encountering an error encoding metric attributes in the
OTLP exporter, log the key that had an error.
* Fix `ExponentialHistogramAggregation`
* Log a warning when a `LogRecord` in `sdk/log` has dropped
attributes due to reaching limits
* Fix RandomIdGenerator can generate invalid Span/Trace Ids
* Add Python 3.12 to tox
* Improve resource field structure for LogRecords
* Update Semantic Conventions code generation scripts: - fix
namespace exclusion that resulted in dropping `os` and `net`
namespaces. - add `Final` decorator to constants to prevent
collisions - enable mypy and fix detected issues - allow to
drop specific attributes in preparation for Semantic
Conventions v1.26.0
* Update semantic conventions to version 1.26.0.
* Use semconv exception attributes for record exceptions in
spans
* Fix _encode_events assumes events.attributes.dropped exists
* Validate links at span creation
* Add attributes field in `MeterProvider.get_meter` and
`InstrumentationScope`
* Fix inaccessible `SCHEMA_URL` constants in `opentelemetry-
semantic-conventions`
-------------------------------------------------------------------
Mon Jun 17 11:39:29 UTC 2024 - Dirk Müller <dmueller@suse.com>
- update to 1.25.0:
* Fix class BoundedAttributes to have RLock rather than Lock
* Remove thread lock by loading RuntimeContext explicitly.
* Update proto version to v1.2.0
* Add to_json method to ExponentialHistogram
* Bump mypy to 1.9.0
* Fix exponential histograms
* Fix otlp exporter to export log_record.observed_timestamp
* Add capture the fully qualified type name for raised
exceptions in spans
* Prometheus exporter sort label keys to prevent duplicate
metrics when user input changes order
* Rename test objects to avoid pytest warnings
* Add span flags to OTLP spans and links
* Record links with invalid SpanContext if either attributes or
TraceState are not empty
* Add OpenTelemetry trove classifiers to PyPI packages ([#3913]
* Fix prometheus metric name and unit conversion (#3924)
* Add parameters for Distros and configurators to configure
autoinstrumentation in addition to existing environment
variables.
-------------------------------------------------------------------
Wed Apr 3 07:25:46 UTC 2024 - Dirk Müller <dmueller@suse.com>
- update to 1.24.0:
* Make create_gauge non-abstract method
* Make `tracer.start_as_current_span()` decorator work with
async functions
* Fix python 3.12 deprecation warning
* bump mypy to 0.982
* Add support for OTEL_SDK_DISABLED environment variable
* Fix ValueError message for PeriodicExportingMetricsReader
* Use `BaseException` instead of `Exception` in
`record_exception`
* Make span.record_exception more robust
* Fix license field in pyproject.toml files
-------------------------------------------------------------------
Sat Mar 23 22:18:01 UTC 2024 - Jaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@fastmail.net>
- update to 1.23.0:
* Use Attribute rather than boundattribute in logrecord (#3567)
* Fix flush error when no LoggerProvider configured for LoggingHandler (#3608)
* Fix OTLPMetricExporter ignores preferred_aggregation property (#3603)
* Logs: set observed_timestamp field (#3565)
* Add missing Resource SchemaURL in OTLP exporters (#3652)
* Fix loglevel warning text (#3566)
* Prometheus Exporter string representation for target_info labels (#3659)
* Logs: ObservedTimestamp field is missing in console exporter output (#3564)
* Fix explicit bucket histogram aggregation (#3429)
* Add code.lineno, code.function and code.filepath to all logs (#3645)
* Add Synchronous Gauge instrument (#3462)
* Drop support for 3.7 (#3668)
* Include key in attribute sequence warning (#3639)
* Upgrade markupsafe, Flask and related dependencies to dev and test
environments (#3609)
* Handle HTTP 2XX responses as successful in OTLP exporters (#3623)
* Improve Resource Detector timeout messaging (#3645)
* Add Proxy classes for logging (#3575)
* Remove dependency on 'backoff' library (#3679)
-------------------------------------------------------------------
Mon Jan 15 21:10:53 UTC 2024 - Dirk Müller <dmueller@suse.com>
- Initial package (1.22.0)
