You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
What is the optimal and suggested way, if even possible, to somehow be able to host configurations for various different backends that could not adhere to the spec completely? In the same process.
For example, I use Sentry, which treats traceId more like a session ID - i.e. multiple requests on the same navigation page are part of the same trace and Sentry has a concept of transaction as an individual part of the trace. Sentry uses sentry-trace as a propagation carrier set through the SDK or manually.
At the same time, I use GCP loadbalancer, which will use the provided W3C traceparent or set one in the logs and ongoing request to the backend. But for those I don't want Sentry behaviour I would like to be able to differentiate otel trace from trace in GCP logs.
Also I could have different OTel-compliant backends like Jaeger, Datadog, whatnot.
How to connect those things so I have kind of duality between processing/propagating context and creating spans for different backend needs and in the same time keep them as close as possible.
Seems like because of how Sentry works I will need a pair of different traceId to use for it and the rest. Does it mean I would also need to have completely different spans?
Is there an example that deals with those issues? Or way of thinking about it?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
Hi!
What is the optimal and suggested way, if even possible, to somehow be able to host configurations for various different backends that could not adhere to the spec completely? In the same process.
For example, I use Sentry, which treats
traceIdmore like a session ID - i.e. multiple requests on the same navigation page are part of the same trace and Sentry has a concept of transaction as an individual part of the trace. Sentry usessentry-traceas a propagation carrier set through the SDK or manually.At the same time, I use GCP loadbalancer, which will use the provided W3C
traceparentor set one in the logs and ongoing request to the backend. But for those I don't want Sentry behaviour I would like to be able to differentiate otel trace from trace in GCP logs.Also I could have different OTel-compliant backends like Jaeger, Datadog, whatnot.
How to connect those things so I have kind of duality between processing/propagating context and creating spans for different backend needs and in the same time keep them as close as possible.
Seems like because of how Sentry works I will need a pair of different
traceIdto use for it and the rest. Does it mean I would also need to have completely different spans?Is there an example that deals with those issues? Or way of thinking about it?
Thanks!
Beta Was this translation helpful? Give feedback.
All reactions