fix(basic-tracer): add timestamp for event

[mayurkale22]

Closes #191

originates from https://github.com/open-telemetry/opentelemetry-js/pull/192/files#diff-08ba7b60ae6ed103fba2d7664cff8c92R88

[codecov-io]

Codecov Report

Merging #195 into master will increase coverage by 0.02%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #195      +/-   ##
==========================================
+ Coverage   93.59%   93.62%   +0.02%     
==========================================
  Files          50       50              
  Lines        1578     1585       +7     
  Branches       99       99              
==========================================
+ Hits         1477     1484       +7     
  Misses        101      101

Impacted Files	Coverage Δ
packages/opentelemetry-basic-tracer/src/Span.ts	92.72% <100%> (ø)	⬆️
...kages/opentelemetry-basic-tracer/test/Span.test.ts	100% <100%> (ø)	⬆️

[draffensperger]

Could we also make an optional time parameter for the event?

That will be helpful for bringing over the OpenCensus Web code since it generates timed events based on browser performance data (this is similar to my desire to make the start span take a start time and end span take an end time, etc.)

[mayurkale22]

This is not defined in the specs: https://github.com/open-telemetry/opentelemetry-specification/blob/master/specification/api-tracing.md#add-events. AFAIK SDK is responsible for adding timestamp internally.

[draffensperger]

Hmm, OK. Should I raise an issue with the specs for this? What's the process for that?

I do need the feature for web tracing.

[mayurkale22]

so the event timestamp is gonna be different from current timestamp (which we are adding internally) in web tracing?

What's the process for that?

Create new issue here and start the discussion.

[draffensperger]

Thanks for the pointer! I've created open-telemetry/opentelemetry-specification#213 to discuss and shared more about the web tracing use case there (basically generating timed events from Performance API data as measured by the browser runtime rather than in the JS code)

[bg451]

Oh huh okay I think that part of the spandata removal is going to have to include optional timestamps for events, because how else would you do out of band reporting without the timestamp? I'll comment on that issue @draffensperger

[mayurkale22]

Thanks for the issue @draffensperger. If all goes well, I will open new PR to merge TimedEvent and Event interface and add optional timestamp param to addEvent method. To unblock #192 work I'd like to merge this one. /cc @hekike

[draffensperger]

LGTM on this given the current spec

[hekike]

Thanks!