#Custom timeout does not send our live tracing to honeycomb

1 messages · Page 1 of 1 (latest)

tawdry quail
#

👋 Folks

We are using dagger sdk/engine 0.13.0. Our tracing looks superb and we also have custom logic for timeout(we call close.client()) when particular job is executing for more than 15mis.
And for this special case, we aren't seing any live traces coming to honeycomb. I am not even sure whether its a live traces or traces sent after full completion.

Could you please guide?

elfin axle
#

honeycomb doesn't support live tracing

#

live traces can only be supported in dagger cloud as i'm aware - other tracing stores just don't support it as a feature

tawdry quail
#

@elfin axle can you guide on how to manually flush dagger traces from application itself, this is needed when our custom timeout logic is executed, we call os.exit(1) internally and we do cleanup before that, but dagger traces I believe we also need to do flush

quartz verge
# tawdry quail <@488718750690967563> can you guide on how to manually flush dagger traces from ...

the dagger client automatically waits for all OTEL data to be flushed before finishing: https://github.com/dagger/dagger/blob/7e774f69dd089714531b7f45df3596a7c9501951/engine/client/client.go?plain=1#L555

GitHub

An engine to run your pipelines in containers. Contribute to dagger/dagger development by creating an account on GitHub.