feat(telemetry): report export status, duration, and phase timings

[FarhanAliRaza]

Track per-phase timings (compile/build/zip) and success/failure for reflex export, and stop mutating the cached event defaults so kwargs from one event can't leak into the next.

All Submissions:

• Have you followed the guidelines stated in CONTRIBUTING.md file?
• Have you checked to ensure there aren't any other open Pull Requests for the desired changed?

Type of change

• New feature (non-breaking change which adds functionality)

New Feature Submission:

• Does your submission pass the tests?
• Have you linted your code locally prior to submission?

Changes To Core Features:

• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your core changes, as applicable?
• Have you successfully ran tests with your changes locally?

fixes ENG-9477

[codspeed-hq]

Merging this PR will not alter performance

✅ 24 untouched benchmarks

---

_{Comparing FarhanAliRaza:export-event (e7711f7) with main (9ed3692)}

[greptile-apps]

Greptile Summary

This PR adds per-phase timing (compile_duration, build_duration, zip_duration), overall duration, and status/detail fields to the reflex export telemetry event, and fixes a long-standing bug where mutating the cached get_event_defaults() properties dict caused kwargs from one _send call to leak into subsequent calls.

• export.py: Wraps each phase in a _time_phase closure-based context manager that records duration into phase_durations via a finally block, ensuring partial timings are captured even on failure; telemetry is fired from a top-level finally so it is always sent.
• telemetry.py: Adds the new fields to additional_keys, switches from properties = event_data["properties"] (mutating shared reference) to .copy(), and tightens the filter to drop None-valued kwargs rather than propagating them.
• Tests: Existing test_send is expanded with parametrized cases; four new focused tests cover no-leak isolation, unknown-kwarg dropping, None-value filtering, and the new test_export.py covers success, all four partial-failure points, backend-only, and no-zip scenarios.

Confidence Score: 4/5

Safe to merge — the core bug fix is correct and the new telemetry instrumentation is well-tested across all branch paths.

The cached-mutation fix is correct and thoroughly exercised. The two observations (setup_frontend grouped into build_duration, overwrite guard removed) are non-breaking design choices with no effect on production behaviour today.

reflex/utils/export.py is worth a second glance on the phase grouping decision if PostHog analysis needs to distinguish frontend-setup failures from build failures.

Important Files Changed

Filename	Overview
reflex/utils/export.py	Wraps compilation, build, and zip steps in per-phase timers; reports status, detail, and durations to telemetry in a finally block.
reflex/utils/telemetry.py	Fixes the cached-mutation bug by copying properties before modifying; expands additional_keys with new export fields; new logic correctly filters None values.
tests/units/test_telemetry.py	Mock target correctly changed from _get_event_defaults to get_event_defaults; new parametrized and dedicated isolation/None-filter tests added.
tests/units/utils/test_export.py	New test file covering success, all four partial-failure points, backend-only, and no-zip scenarios.

Sequence Diagram

sequenceDiagram
    participant C as Caller
    participant E as export()
    participant P as prerequisites
    participant B as build
    participant T as telemetry.send()
    participant PH as PostHog API

    C->>E: export(frontend, zipping, ...)
    E->>E: "start = time.monotonic()"
    E->>E: "phase_durations = {}"
    E->>E: "status = success"

    opt "frontend=True"
        E->>P: get_compiled_app() [_time_phase(compile_duration)]
        P-->>E: ok / raises
        E->>B: setup_frontend() + build.build() [_time_phase(build_duration)]
        B-->>E: ok / raises
    end

    opt "zipping=True"
        E->>B: zip_app() [_time_phase(zip_duration)]
        B-->>E: ok / raises
    end

    alt Exception raised
        E->>E: "status=failure, detail=type(exc).__name__, raise"
    end

    E->>T: send(export, status, detail, duration, compile_duration, build_duration, zip_duration)
    T->>T: _prepare_event(): copy defaults, filter None kwargs
    T->>PH: POST event (async background task)
    T-->>E: done (fire-and-forget)
    E-->>C: return (or re-raise)

Loading

_{Reviews (1): Last reviewed commit: "feat(telemetry): report export status, d..." | Re-trigger Greptile}

[masenf]

this should be a fixture so it doesn't have to be called in every test case that wants to use it

[masenf]

this should probably also be a fixture as well. these set up steps shouldn't comingle with the test logic. AI likes to do that a lot.