Conversation
size-limit report 📦
|
JPeer264
left a comment
There was a problem hiding this comment.
Tbh I don't feel quite comfy to approve this. We get rid of a lot of assertions. And the flaked test actually has one extra trace with the "status": "internal_error",
So not sure if it is our goal to make it less flaky by just making it more loose. What would happen if we accidentally add another span (e.g. double instrumentation), this would get lost with it, as we would never know
chargome
left a comment
There was a problem hiding this comment.
LGTM, but we recently talked about being more conservative with arrayContaining and objectContaining and instead just assert more specifically
|
Hmm I am not 100% sure. IMHO the stuff I dropped in this PR are not really important - it's just what otel emits today, but if we were to re-implement this we likely would not care about these. So not sure if it makes sense to assert that they are all there if we don't really care about them 🤔 If we care about those, then yes we should assert that they are in there (but still do so in a way that does not assert on the array order). If we do not omit these, then we need to somehow rewrite this test to figure out when it is adding these spans and when not (as it is apparently not that deterministic as of now as they are usually not there but apparently sometimes there). Not sure I agree with the sentiment that we should check for exact array matches, as this kind of ends up testing things that we do not actually care to test. e.g. what if some other thing emits a span into this, or the order of spans changes, these are all things that do not relate to this test at all so we should not assert on it specifically IMHO. I also do not actually want to spend too many cycles on this though, so I'll just close this for now and we can always revisit it :) |
Closes #20424