File python-prometheus-client.changes of Package python-prometheus-client
------------------------------------------------------------------- Thu Nov 14 08:15:20 UTC 2019 - Thomas Bechtold <tbechtold@suse.com> update to version v0.7.1 * Fix lint related to operator positioning (#273) * Fix file descriptor not closed (#269) * Add basic openmetrics exposition * Add gsum/gcount to GaugeHistogram. * Fix multi process collision by initializing pid variable in master process (#192) * Use less genexprs in multiprocess accumulate * Refactor MetricWrapper magic to class hierarchy * Prevent double reading all values when collect multiprocess metrics (#262) * Fix lint around unused and star imports (#245) * Optimizations for tuples using. Removed unused types transformations. (#263) * Read only `used` bytes from MmapedDict files, not all the zeroes too * Check that quantile and le labels have Go %g syntax, with .0 (#346) * Run isort * Avoid unpack_from() for a simple slice * Fix deadlock on gcCollector (#371) * Fix: #222. Support custom registry for MetricsHandler. (#223) * Enable flake8 import lint checking (#327) * Avoid duplicate JSON parsing and small allocations * correct spelling mistake (#193) * Fix unescaping (#291) * Add the GaugeHistogram * Remove unused time import (#200) * Disable GCCollector in multiprocess mode (#324) * Make WSGI metrics server not log requests (#203) * Add test for empty metadata, and the null byte. * helpful error message when multiprocessing dir env is unset (#206) * Add _created to Counter/Summary/Histogram * Add Info metric type. * Simplify code a bit * Add support for int values, UNIT, more tests. * Use `collections.defaultdict` to speed up metrics counting (#257) * Add UntypedMetricFamily to prometheus_client.core (#220) * Thread safety done right (#290) * Fix lint related to vertical whitespace (#246) * parser: ensure samples are of type Sample (#358) * Reduce unnecessary metric creation in loop (#256) * Remove duplicate code (#320) * Only test twisted on 3.7 (latest minor) (#428) * Fix thread leak in Python 3.7 #340 (#356) * b"Check exemplars are only where they're meant to be." * Add support for parsing timestamps * Fix #95: catch OSError when collecting file descriptors. (#205) * b"multiprocess: don't crash on missing gauge_live/sum files (#424)" * PEP333 does not require QUERY_STRING to be present. (#252) * Also set the mmap mode to match the read_mode flag. * Construct less tuples and dicts in accumulate * b"Avoid re-entrant calls to GC collector's callback (#343)" * Add python3.7 support (#418) * b"Check samples are grouped, untyped isn't used, and for invalid/missing le/quantile values." * b"Don't use mmap() when only reading a MmapedDict file" * Add OM timestamp support, and unittests * Fix lint for line continuation indents (#242) * Check for NaN counter values * Multiprocess mode: using direct assignment for writing into mmap (#315) * If only reading the mmap file, open in read-only mode. * Inirial work on OpenMetrics parser. * In multiprocess mode, ensure that metrics initialise to the correct (#328) * Example on how to expose metrics in a Flask application (#297) * Check for invalid info/stateset values. * Fix lint around operator spacing (#244) * Test and support Python 3.6 (#181) * py26 has fixed dependencies (#218) * Catch repeated metadata, and metadata after samples. * Added metrics to MetricFamily exceptions to fix #362 (#364) * Factor out value/timestamp/label parsing * Remove travis pypy version hack (#240) * optimize openmetrics text parsing (~4x perf) (#402) * Text parser optimization (~4.5x perf) (#282) * Check bucket rules are being followed * Permit subclassing of MetricsHandler (#339) * Document removal of _total from metric name. * Use precompiled struct methods for pack, unpack data (#266) * Add detailed Garbage Collection statistics. (#301) * Set default timeout of pushgateway actions to 30s (#237) * Fix bug in WSGI app code. * Add support for modelling and exposing exemplars * Add support for parsing exemplars * Add some more todos * Add tox env for running flake8 lint (#238) * Refactor everything out from `core.py` (but keep it for re-exports) * Update OM content type. * Check for negative counter-like, guage histogram, and quantile values. (#338) * Code style fixups (#325) * Disable gcCollector for pypy (#380) * Split MultiProcessCollector.__init__ for better profiling * Refactor MultiProcessCollector.collect() to allow for arbitrary merging. (#302) * read stat and limits files in binary mode (#234) (#235) * Ensure MetricFamily labels is a tuple (#184) * Fix SyntaxWarning due to unescaped sequence (#366) * Add support for Enum/StateSet * Change exemplar length limit to be only for label names+values (#397) * Fix small spelling error in README.md (#417) * Fix import of openmetrics, and make accept handling more general. (#318) * Add timestamp type for nano resolution * Check ordering within groups. * Check for non-real timestamps * Improve tests, handle trailing space after value. * Limit memory reads for mmaped files (#329) * Make indentation multiple of 4 spaces (#241) * Fix F811 redefinition lint (#239) * fstat mmap file only once * Expand unit support. * check Python version with sys.version_info (#227) * Fix issue #208: parse the correct scheme of URL in Python26 (#209) * Drop samples with duplicate timestamps, presumably due to truncation * compact MethodType between 2 and 3 (#403) * Make the format strings compatible with Python 2.6 (#361) * Check for exemplars being too long. * Switch to a namedtuple Sample for samples. * Make counter metric name not have _total internally. * flake8 compliance (#331) * flake8 test passing (#373) * Fix instrumentation link in README (#384) * Fix lint related to assigned but unused variables (#274) * multiprocess: pass None as default argument (#212) * Update CONTRIBUTING.md for DCO, test, and lint (#326) * OM changes untyped to unknown - Drop python-devel BuildRequires. It's not needed ------------------------------------------------------------------- Mon Jul 31 15:00:02 UTC 2017 - vtheile@suse.com Initial package (0.0.20) for openSUSE.
