Summary

Move the "end-only span ingest" optimization (originally done on the EGP backend in scaleapi#145863) into the SDK producer, so the wasted span-start write is never sent on the wire in the first place.

Tracing writes each span twice — once on start (no end_time) and once on end. The start row is only ever overwritten by the end write moments later, so persisting it:

• doubles span-ingest write volume (HTTP + DB), and
• on the SGP backend costs a non-HOT UPDATE per span (the joined_data_tsvector GIN index is recomputed when output arrives on the end write) plus a dead tuple.

This change makes SGPSyncTracingProcessor / SGPAsyncTracingProcessor skip the span-start upsert, so each span is persisted once, on end (a single INSERT server-side). Doing it in the SDK also eliminates the wasted start HTTP call entirely — not just the DB write.

Behavior

• Default ON (skip span-start). on_span_start (sync) and on_spans_start (async) become no-ops by default.
• Set AGENTEX_TRACING_SKIP_SPAN_START=0 (false/no/off also work) to restore the start write — e.g. if you need in-flight spans visible before they complete, or spans that never end (process crash) to still be persisted.
• on_span_end / on_spans_end are unchanged; the end write already carries the full span (start_time + end_time + input/output), so nothing is lost in the default path for spans that complete normally.

Parsing mirrors the SDK's existing AGENTEX_TRACING_METRICS convention (raw not in ("0","false","no","off")).

Trade-offs

• In-flight spans are not visible until they complete.
• Spans that never end (e.g. a crash before on_span_end) are not persisted.

Both are reversible per-deployment via the env var.

Tests

• New: on_span_start/on_spans_start are no-ops by default; emit when skip is disabled; _skip_span_start_enabled() env parsing (default + falsy/other values).
• Updated the existing two-write lifecycle / start-batching / start-metrics tests to set AGENTEX_TRACING_SKIP_SPAN_START=0 so they continue to exercise the start path.
• All 31 tests in test_sgp_tracing_processor.py pass; ruff check clean.

Note

Branched off v0.11.8 per request; rebased onto main (0.12.0) for a clean PR — the processor file and its tests are byte-identical between the tag and main, so the code change is exactly the same.

🤖 Generated with Claude Code

Greptile Summary

This PR moves the "end-only span ingest" optimisation from the EGP backend into the SDK producer itself, eliminating the wasted span-start HTTP call entirely. SGPSyncTracingProcessor.on_span_start and SGPAsyncTracingProcessor.on_spans_start become no-ops by default, controlled by a new AGENTEX_TRACING_SKIP_SPAN_START env var that mirrors the existing AGENTEX_TRACING_METRICS convention.

• sgp_tracing_processor.py: Adds _skip_span_start_enabled() (reads env var dynamically on every call), gates both start handlers behind it, and logs the active mode at processor construction time.
• Test suite: Adds dedicated tests for default-skip and opt-in-emit paths and TestSkipSpanStartEnv env-var parsing; backfills AGENTEX_TRACING_SKIP_SPAN_START=0 on the four existing tests that exercise the start write path, preserving their semantics.

Confidence Score: 5/5

Safe to merge. The change makes span-start writes a no-op by default with a clearly documented opt-out; on_span_end is untouched and carries the full span payload, so no trace data is lost for normally-completing spans.

The logic is straightforward: a single env-var gate in each start handler, all end-path code unchanged. The new TestSkipSpanStartEnv class directly exercises every falsy/truthy value, and the four updated existing tests correctly re-enable the start path via monkeypatch.setenv. No span data is dropped for spans that complete normally.

No files require special attention.

Important Files Changed

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[Span Start Event] --> B{_skip_span_start_enabled?\nenv: AGENTEX_TRACING_SKIP_SPAN_START}
    B -- "True (default)" --> C[No-op: return immediately\nNo HTTP call sent]
    B -- "False (opt-in)" --> D[Build SGPSpan\n_build_sgp_span]
    D --> E[flush / upsert_batch\nSGP Backend HTTP write]
    F[Span End Event] --> G[Build SGPSpan + end_time\n_build_sgp_span]
    G --> H[flush / upsert_batch\nSGP Backend HTTP write\nSingle INSERT on backend]

Reviews (4): Last reviewed commit: "refactor(tracing): split span batch/ling..." | Re-trigger Greptile