Closed yovelcohen closed 1 week ago
k8se-otel.k8se-apps.svc
is not the Logfire API. If you didn't configure that yourself, then I think something in Azure is setting the OTEL_EXPORTER_OTLP_ENDPOINT
environment variable. You can override it with base_url='https://logfire-api.pydantic.dev'
in logfire.configure()
.
@alexmojaki thanks, that was indeed the issue, I'm just a bit curios as to why would this happens on some container apps and not others, but thanks :)
I see OTEL_EXPORTER_OTLP_ENDPOINT
mentioned in https://learn.microsoft.com/en-us/azure/container-apps/opentelemetry-agents?tabs=arm#environment-variables, but I don't know the details. If you learn more, please share :)
I'm closing since this seems resolved, but if other people run into the same issue then we might want to consider not paying attention to OTEL_EXPORTER_OTLP_ENDPOINT
to avoid confusion.
Description
I have a bunch of python microservices, currently running on azure container apps. they write logs to two major projects in Logfire. For some reason, even though I'm using the same configuration code and the exact same write token, some of the services can't connect to logfire, with a weird open telemetry error msg:
I haven't been able to find consistency as to why some services can connect and some connect (double-checked the logfire configuration in each)
Python, Logfire & OS Versions, related packages (not required)