Bump bundled opentelemetry-proto to v1.10.0 and regenerate#5223
Conversation
|
|
8644435 to
fa4f9b6
Compare
|
Please update the |
Address review feedback on open-telemetry#5223: the bundled proto definitions were bumped from v1.9.0 to v1.10.0, so the README version references should match. Co-Authored-By: Claude Opus 4.7 (1M context) <[email protected]>
|
Thanks for the quick review @herin049 ! Done — updated both v1.9.0 references in I left |
Thanks. I've opened another PR to bump the JSON package to v1.10.0 already. |
|
does this need an |
Update PROTO_REPO_BRANCH_OR_COMMIT in scripts/proto_codegen.sh from v1.9.0 to v1.10.0, regenerate the generated *_pb2.py / *_pb2.pyi files, and update the opentelemetry-proto README version references to match. The motivation is the profiles v1development proto: in the v1.9.0 pin the Sample message carried outdated field numbers (stack_index=1, values=2, attribute_indices=3, link_index=4, timestamps_unix_nano=5). v1.10.0 (2026-03-09) renumbers it to stack_index=1, attribute_indices=2, link_index=3, values=4, timestamps_unix_nano=5, which the regenerated descriptors now reflect. The bump also pulls in unrelated proto changes between v1.9.0 and v1.10.0: common v1 gains AnyValue.string_value_strindex and KeyValue.key_strindex fields (development, profiling-oriented), and a SeverityNumber doc-comment was dropped in logs v1.
31d7886 to
33ea2bc
Compare
Done! |
Description
Bumps the bundled proto definitions from
opentelemetry-protov1.9.0tov1.10.0and regenerates the*_pb2.py/*_pb2.pyifiles viascripts/proto_codegen.sh(onlyPROTO_REPO_BRANCH_OR_COMMITwas edited; nomanual edits to generated files).
The motivating change is in the
v1developmentprofiles signal. Betweenproto
v1.9.0andv1.10.0theSamplemessage field numbers wererenumbered to group the identity fields first:
stack_indexattribute_indiceslink_indexvaluestimestamps_unix_nanoBecause the bundled snapshot still uses the old numbering, a
Sampleserialized via
opentelemetry-protois wire-incompatible with OTLP profilesconsumers built against current proto (e.g. Grafana Pyroscope 2.x):
valuesis decoded as
attribute_indicesand the profile is silently dropped.The bump also pulls in two unrelated
v1.10.0changes:common/v1:AnyValuegainsstring_value_strindexandKeyValuegainskey_strindex(development, profiling-oriented string-table references).logs/v1: a doc comment was dropped fromSEVERITY_NUMBER_UNSPECIFIED(
.pyionly, no wire change).Traces and metrics protos are unchanged.
Type of change
How Has This Been Tested?
scripts/proto_codegen.shruns cleanly and regenerates the bindings. Therenumbered
Sampledescriptor was verified to round-trip against a currentOTLP profiles consumer.
🤖 Generated with Claude Code