Unique Spans#7882
Conversation
Unit Test ResultsSee test report for an extended history of previous test failures. This is useful for diagnosing flaky tests. 20 files ± 0 20 suites ±0 11h 30m 17s ⏱️ - 18m 10s For more details on these failures, see this check. Results for commit b474c16. ± Comparison against base commit 129b7cb. ♻️ This comment has been updated with latest results. |
| ts.annotations.pop("span", None) | ||
|
|
||
| def _ensure_default_span(self) -> Span: | ||
| """Return the currently active default span, or create one if the previous one |
There was a problem hiding this comment.
Out-of-scope: I'm not a huge fan of this as it's not very reproducible. Let's keep thinking about a better way to reproduce the progress/code-tracking behavior from computations.
There was a problem hiding this comment.
I don't know; I don't hate it now.
If you don't explicitly use spans, you get something that somehow makes sense, at the cost of some mild non-determinism.
If you do use spans, everything is deterministic.
2 participants
Change spans so that, if a user opens and reopens a span context manager with the same tag, it counts as separate spans - e.g. in case of repeated calsl to the same high-level library.
The 'default' span now automatically closes and reopens itself into a new 'default' span when the cluster is idle, save for all tasks not explicitly annotated with
span. This effectively replicates 1:1 the behaviour of computations.CC @hendrikmakait @mrocklin @ntabris