trunk-merge/pr-3478/5fe47493-2526-4e60-9434-dc736114dd1c#3770
Closed
trunk-io[bot] wants to merge 21 commits into
Closed
trunk-merge/pr-3478/5fe47493-2526-4e60-9434-dc736114dd1c#3770trunk-io[bot] wants to merge 21 commits into
trunk-io[bot] wants to merge 21 commits into
Conversation
…OTLP Adds OtelRunTelemetry, a SessionLogWriter sink that ships an allowlisted metadata subset of the session log (run/turn/tool lifecycle, usage, errors; never message content or tool arguments) to PostHog Logs, and, when a traces URL is configured, builds one APM trace per run (task_run root span, a turn span per prompt, a tool_call:<kind> span per tool call) with trace/span ids stamped on log records so Logs and APM cross-link. Resource attributes carry run_id/task_id/team_id/user_id/distinct_id so cloud runs are filterable per user in the Logs UI. Configured via POSTHOG_AGENT_OTEL_LOGS_URL/_TOKEN (+ optional POSTHOG_AGENT_OTEL_TRACES_URL), deliberately not standard OTEL_* names so OTel SDKs in user code running in the sandbox never auto-export into the telemetry project. Telemetry stays off unless the logs pair is set, flushes on session cleanup/terminal errors, and can never break session log persistence (sink failures are isolated). Replaces the unwired otel-log-writer whose default endpoint (/i/v1/agent-logs) does not exist in the ingest service, and removes the dead AgentConfig.otelTransport field. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
…emetry Addresses three review findings: - The sandbox never emits task_complete for successful runs (the terminal "completed" status is decided by the workflow outside), so the root span was ending with unset status on success. The trace builder now marks the root span OK when a turn ends cleanly with end_turn; a run error still always wins and a turn completion arriving after an error cannot flip the status back. - reportFatalError (uncaught exception / unhandled rejection) marked the run failed via the API but never touched telemetry, leaving hard crashes invisible in the new pipeline. It now mirrors an error record (error_source=agent_server_crash) and shuts telemetry down, ending the root span as errored and flushing before the process dies. - The README "Writing logs" section still described the removed otel-log-writer as the preferred full-transcript path to the nonexistent /i/v1/agent-logs endpoint; rewritten to match the sink-based metadata export. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
…est turn Addresses two external review findings: - _posthog/console records are no longer exported to PostHog Logs. Console lines are free-text agent-server diagnostics that interpolate arbitrary data - the user-message handler logs a 100-char prompt preview and the extension-notification handler stringifies params into that channel - so exporting them verbatim violated the metadata-only allowlist. They remain in the S3 session log and the event-ingest stream. Regression tests assert console entries (including an interpolated prompt preview) never map to an exported record. - The task_run root span status is now resolved at shutdown from the LATEST turn outcome instead of being set sticky-OK on the first clean end_turn. A multi-turn run whose final turn is cancelled/refused no longer reports OK; a run error still always wins. The root span only exports at end, so per-turn status writes were cosmetic anyway. Stacked on posthog-code/agent-run-otel-telemetry (pure fast-forward). Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
… fixes Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Addresses four review findings on the run telemetry export: - _posthog/error records and the root span no longer carry the raw error string: it is free text that can embed prompt or repo content (exception paths, provider errors). Exported signal is now the generic "run error" body plus error_source/stop_reason attributes; the full message stays in the session log and the task run's error_message. - Telemetry flush/shutdown use Promise.allSettled so logs and traces complete independently, and both batch processors cap exports at 5s (exportTimeoutMillis) instead of the SDK's 30s default - a rejecting or hanging traces endpoint can no longer starve log delivery or delay session cleanup. - A run error now closes still-open tool spans as ERROR with tool_status=interrupted, so APM never shows a healthy-looking active tool under a failed run. - OtelTransportConfig/AgentConfig.otelTransport are restored as @deprecated ignored stubs: @posthog/agent is published, so removing exported types is an API break reserved for a major. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
The task detail routes redirected away on any 404 from the detail query, even when a cached or list copy of the task exists. Optimistic and cloud-pending tasks are not returnable by the API yet (see the route loader comments), so the unconditional redirect kicked users off tasks they had just created. The detail query still always runs (so a stale cached copy converges on the server's latest run state), but a 404/missing result only triggers the redirect when there is no usable cached task to render. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Local verification of the telemetry branch showed logs and the turn span landing in ClickHouse but never the task_run root span, even after the run reached completed and extra delay. Root cause: the root span only ended in cleanupSession, which is only reachable via SIGTERM or the close command - and sandbox teardown never delivers SIGTERM to agent-server, because it is an exec'd process (docker stop signals only the container's PID 1; Modal terminate is immediate). The root span was still open when the process was killed, so it never exported. The turn span survived only because it ended mid-run and the 2s batch exporter got it out. Fix: end telemetry eagerly at the run's in-process terminal points instead of relying on teardown: - finalizeRunTelemetry: full telemetry shutdown (ends the root span with the resolved status and drains both exporters) when a background run's initial or resume prompt settles - the run is over in-sandbox at that point; the workflow marks the terminal status and destroys the sandbox right after. - signalTaskComplete: upgrade the terminal-failure flush to a full shutdown, after the error mirror is appended, so failed runs export a root span with ERROR status. - cleanupSession remains the path for interactive close. Known limitation, now documented: an interactive session ended by hard teardown (e.g. inactivity timeout) loses the root span; its turn/tool spans and logs still assemble under the same trace id. Tests: order assertion that the error mirror lands before the terminal shutdown, and a parameterized check that finalizeRunTelemetry fires for background runs only. The full agent-server + telemetry suites pass (161). Stacked on posthog-code/agent-run-otel-telemetry (pure fast-forward). Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Rebasing the terminal-state hardening onto main's evolved sessionService adopted main's cloudHydrationPromises in-flight memoization in place of the original hydration-key set. Keying that map by run id alone reintroduced the bug the original fix addressed: a run going terminal while a resume-chain (or single-run) hydration was still in flight would dedupe onto the pending promise and never apply the final transcript. The map is now keyed by run id + hydration mode (terminal-chain / resume-chain / single). Also updates the two terminal-hydration host tests to main's getTaskRunSessionLogsResult client API (result-object shape, and the ancestor+current parallel fetch that makes each resume hydration two calls). Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Mirrors the PostHog/posthog report update: documents the incorporated scout run log mirror (#71094) alongside the OTLP run-telemetry export, including the two pipelines' differing privacy postures and scopes. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Mirrors the PostHog/posthog report update: the scout run log mirror gained a direct OTLP delivery leg (TASK_RUN_LOGS_MIRROR_OTLP_URL/_TOKEN) for hosts whose stdout no collector tails - in local dev append_log runs in the host Django process, which the dev collector (docker containers only) never sees. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Review finding: cloudHydrationPromises memoizes terminal-chain and resume-chain hydrations under different mode keys, so both can be in flight for the same run. If the run settles while a resume-chain hydration's fetches are pending, the terminal-chain hydration records the full chain as processed - and the late resume-chain write then clobbered processedLineCount with its smaller leaf-only cursor (updateSession is a blind assign with no monotonic guard). Impact was bounded (the transcript content is identical and the next terminal hydration self-heals), but it violated the invariant the mode-keying fix set out to guarantee: the terminal transcript is authoritative. Re-read the session at write time and floor a non-terminal cursor at the settled value once cloudStatus is terminal. The regression test resolves the resume-chain fetches after the run settles and asserts the settled cursor survives; verified to fail without the fix. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
…state Findings from a 4-angle simplify pass (reuse / simplification / efficiency / altitude), no behavior change except where noted: - Extract resolveEffectiveCloudStatus: the 9-line terminal-wins status precedence chain was duplicated byte-for-byte in deriveCloudRunState and deriveSessionViewState; three review angles flagged it independently. One exported helper in cloudRunState.ts now owns the policy. - isTaskDetailNotFoundError uses requestErrorStatus(error) === 404 instead of string-matching "[404]" in the message - the fetcher keeps that message format only as legacy, and the substring check could false-positive on a non-404 error quoting an upstream body. Behavior tightened accordingly (plain Errors mentioning [404] no longer read as 404s). - taskDetailQuery no longer retries 404s: the detail fetch is now always-on, and optimistic/cloud-pending tasks 404 by design, so react-query's default 3 retries turned every mount of a fresh task into 4 requests. - pickFreshestTask is binary (all six call sites were), which also deletes the logically dead ?? fallbacks at the two hook call sites. - sessionService: merge the duplicated optimistic-clear blocks and skip-apply early returns in performCloudTaskSessionHydration; shrink finalizeTerminalCloudTask to the (taskRunId, status) both call sites use; clearTerminalCloudPromptState derives taskId from the session (removing the dual-key lookup hazard) and now also serves the pre-existing inline terminal cleanup in handleCloudTaskUpdate, so the composer-settling invariant lives in one place; staleness is computed once per update in handleCloudTaskUpdate. - OTel modules: normalizeMethod (the safety-relevant __posthog/ double-prefix rule), asString, and the batch flush-interval constant move to otel-attributes.ts instead of being duplicated across otel-telemetry.ts and otel-trace-builder.ts. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
REPORT.md was a working artifact for the agent session, not part of the change itself. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
…g it The sandbox delivers POSTHOG_AGENT_OTEL_LOGS_TOKEN via the container environment, which agent-spawned tool subprocesses inherit - so any `env` dump in a persisted transcript or PR body would leak the key (flagged by review bots on PostHog/posthog#71100). bin.ts is the only consumer: parse it into the server config, then delete it from process.env so children never see it. Defense in depth, not a boundary - same-UID processes can still read the container's initial env via /proc; the key itself remains a write-only capture token scoped to a dedicated telemetry project. No test: a one-line env mutation at the CLI entry with no cheap seam; exercising it would mean spawning the CLI end to end. Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
Main's #3744 (avoid redundant terminal history fetches) made a terminal resume-chain hydration issue one whole-chain fetch instead of an ancestor+current pair, so the expected call count after the terminal watch drops from 4 to 3. The regression the test pins is unchanged: the terminal hydration must start its own fetch rather than dedupe onto the in-flight resume-chain hydration (count would stay 2 if mode-keying broke). Generated-By: PostHog Code Task-Id: 0c511836-2180-455a-9b58-45df2a0661ec
|
React Doctor found no issues in the changed files. 🎉 Reviewed by React Doctor for commit |
trunk-io
Bot
deleted the
trunk-merge/pr-3478/5fe47493-2526-4e60-9434-dc736114dd1c
branch
July 23, 2026 18:17
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request was created and is being managed by Trunk Merge.
This pull request is based on the main branch at SHA c9605b504f6c9b4ddaab102fc245fd4f5c17827f.
See more details here.
When CI completes, this pull request will be closed automatically.
Pull Requests Being Tested
This pull request is testing the changes from pull request 3478.