OTel traces stuck on "We're still processing this trace" in waterfall view

[sparrowhe]

Self-Hosted Version

26.5.0

CPU Architecture

x86_64

Docker Version

containerd 2.2.3-k3s1

Docker Compose Version

N/A, using Helm Chart

Machine Specification

• My system meets the minimum system requirements of Sentry

Installation Type

Upgraded from 26.4.0 to 26.5.0

Steps to Reproduce

1. Send OpenTelemetry traces from a backend application (in this case, we are using sentry-go)
2. Go to Sentry UI -> Explorer -> Traces.
3. The traces are successfully received and visible in the traces list.
4. Click on any span from the list to view the trace waterfall.
5. The waterfall view fails to load and persistently shows the message: "We're still processing this trace. Please try refreshing after a minute."

Expected Result

When clicking on a span in the Explorer -> Traces list, the trace waterfall view should render correctly and display all associated OpenTelemetry spans.

Actual Result

The waterfall view fails to load and consistently displays the following message:

We're still processing this trace. Please try refreshing after a minute.

This state is permanent. Even after waiting for several hours, the message remains and the waterfall never renders.

Event ID

No response

[ImmortalLotus]

This is also ocurring with Sentry Python SDK. I imagine its a backend issue.

Additionally, after a while on my instance I got this message for a trace that was not opening.

We were unable to find any spans for this trace. If you came here from Logs or Application Metrics, traces may use different sampling rules.

Find similar traces in Explore using [Cross Event Querying](https://docs.sentry.io/product/explore/trace-explorer/#cross-event-querying-beta).

This is a very significant issue, as I basically don't have any tracing at the moment.

Logs still work, and so does alerting . It seems this also affects alerting, as no issues get created just via the logs

All in all this means I have no alerting and no traces

[Gazzonyx]

+1. I"m seeing this with Otel traces not being connected to the parent with the same error message @ImmortalLotus noted. The regression seems to have been introduced in the last minor relase (26.5.0) release cut. This is according to my hazy recollection and gut - not any actual objective data or reproducible steps. So, take it with a grain of salt other than to say, OTel traces hit this same bug, but logs and AI traces are fine.

[aldy505]

If what I said here (#4305 (comment)) is correct, I will mark this issue as duplicate, and we should move there.