fix(docker): langfuse-web healthcheck probes container IP (was falsely unhealthy) - #89
Merged
Merged
Conversation
…calhost langfuse-web (Next.js) binds the container's own IP (e.g. 172.18.0.4:3000), NOT loopback, so the healthcheck's `wget http://localhost:3000/...` was always refused and the container showed perpetually "unhealthy" even though it served fine (host :3100 returned 200 throughout). This misled a data-pipeline investigation into blaming web for thin ingestion. Fix: probe `http://$(hostname -i):3000` via CMD-SHELL. Verified: container goes healthy; reads + OTLP ingestion + score push all work. Also record the 2026-07-30 trace-review data-accrual verification pass in tasks/VERIFY-WHEN-DATA-ACCRUES.md: mark the enrichment-hook-writes-rows + score-push-path items VERIFIED (113 sidecar rows / 24 days; a tessl-push landed in ~5s), and note that the remaining accrual items (score-down flag, rung-3 coverage) are blocked on real volume, not a broken pipe. Separately fixed this pass (runtime only, no source change): removed a redundant crash-looping bundled-hindsight container (the native daemon owns :8888). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
langfuse-webshowedUp (unhealthy)for days, which derailed a trace-review data investigation into blaming it for thin ingestion. It was a false negative: the container serves fine (host:3100returned 200 the whole time; version 3.20). The Next.js server binds the container's own IP (172.18.0.4:3000), NOT loopback, so the healthcheckwget http://localhost:3000/...(and127.0.0.1,[::1]) was always refused.Fix
Probe the container IP via
$(hostname -i)(CMD-SHELL form). Verified: container reportshealthy, and reads + OTLP ingestion + a livetessl-pushscore all work.Also in this PR
tasks/VERIFY-WHEN-DATA-ACCRUES.md: enrichment-hook-writes-rows and score-push-path move to Verified (113 sidecar rows over 24 days; atessl-pushlanded in ClickHouse in ~5s). The remaining accrual items (score-down flag, rung-3 coverage) are blocked on real volume, not a broken pipe —0 scores since 07-06= no create/optimize-skill reviews ran, not dropped pushes.Fixed this session, runtime-only (no source change)
A redundant bundled-hindsight container was crash-looping (missing
ANTHROPIC_API_KEY, and :8888 is owned by the nativecom.user.hindsight-embeddaemon). Removed the orphan;devflow up's guard correctly skips it when the native daemon serves :8888.Verification
make testgreen. Healthcheck confirmed on the running stack (healthyafter recreate). No skill/plugin files touched.