open-telemetry / open-telemetry/opentelemetry-python
Implement higher precision clock than time_ns
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 2.6k
- Forks
- 1k
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 19
Description
Followup to https://github.com/open-telemetry/opentelemetry-python/issues/1594
Currently, we use time_ns() to denote current time in ns -> like here https://github.com/open-telemetry/opentelemetry-python/blob/3ae113816d151af3320b8ce17816a91ebd20e44e/opentelemetry-sdk/src/opentelemetry/sdk/_events/__init__.py#L54
According to Pep0564, the precision varies based off of OS (with testing the average step is that of ~1ms in Windows). Java has implemented an AnchoredClock class which provides much more precise readings of ns (100~ns). We might want to do something similar in our SDKs.
Relevant issue: https://github.com/open-telemetry/semantic-conventions/issues/1621#issuecomment-2552335590
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the time_ns() usage in opentelemetry-sdk/src/opentelemetry/sdk/_events/init.py and read PEP 564's Windows precision notes. Compare the proposed approach with Java's AnchoredClock and the linked semantic-conventions discussion. Done means the Python SDK has an agreed higher-precision clock design and corresponding coverage across the affected SDK usage.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- observability-sre, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100