Docker healthcheck cache or propagation drift response

When this happens

Use this when the panel says fixed but some users still see the old error. Do not stop at the screen message; validate browser cache, CDN cache, DNS TTL, panel propagation, and stale application cache first.

Symptom checklist

• healthcheck appears repeatedly in the Docker UI or logs.
• logs, exit code, volume, network, port bind differs between successful and failed requests.
• The issue appears only after separating image issue from runtime environment issue.
• It often follows deploys, permission changes, configuration edits, or data refreshes.

Likely causes

• healthcheck specifically changes the investigation surface for Docker: verify the exact failing object, route, user, and timestamp before applying the broader pattern.
• The health check exercises a different dependency or path than real users.
• Metrics or logs are missing labels needed to separate real incidents from noise.
• Alert thresholds fire on normal batch or deploy behavior.
• A container restarts before readiness or dependency warmup completes.
• SLO burn is visible before any single log line looks severe.
• For the cache or propagation drift case, the first useful clue is browser cache, CDN cache, DNS TTL, panel propagation, and stale application cache.

First 1-minute checks

1. Write down the first failure time, latest change, affected user, path, and object ID.
2. Compare logs, exit code, volume, network, port bind for success and failure in the same window.
3. Test the hypothesis: operations failure caused by failing health checks, missing telemetry, noisy alert thresholds, deployment loop, or SLO burn.
4. Classify this as cache or propagation drift: browser cache, CDN cache, DNS TTL, panel propagation, and stale application cache.
5. Capture current values before changing configuration.

First evidence

Treat healthcheck as a cache or propagation drift case. First collect evidence for browser cache, CDN cache, DNS TTL, panel propagation, and stale application cache.

Output examples

Normal output

Fresh requests and DNS answers show the same intended value from multiple networks.

Failing output

Only one resolver, browser, CDN edge, or cached object still returns the old value.

Output-to-action branches

• The panel says fixed but some users still see the old error.
  Separate DNS, CDN, browser, and application cache, then purge only the stale layer.
• The working and failing outputs differ.
  Act on the differing layer first: For healthcheck, apply the fix only after reproducing the same condition and saving the before/after evidence for this exact code.
• Command output is normal but users still fail.
  Separate browser cache, cookies, permissions, and network location before declaring it fixed.

Do not do this

• Do not repeatedly edit correct DNS records just because a stale resolver is still visible.
• Do not change multiple layers before identifying the failing layer.
• Do not delete production data, grant broad permissions, or disable security controls as a first response.

Evidence quality

Auto-generated operator draft: includes issue-specific causes, commands, output branches, and unsafe-action warnings. Official-source links and real incident validation are queued for enrichment.

Commands to run first

grep -R "healthcheck" ./logs

docker ps -a

docker logs CONTAINER --tail 200

docker inspect CONTAINER

docker compose config

kubectl get pods -A || true

kubectl describe pod POD || true

grep -R "readiness\|healthcheck\|alert\|latency\|SLO" ./logs

curl -I https://example.com -H 'Cache-Control: no-cache' && dig +trace example.com

Fix order

1. Record the full healthcheck message, failing URL, user, object ID, and latest change.
2. Collect issue-specific evidence for operations failure caused by failing health checks, missing telemetry, noisy alert thresholds, deployment loop, or SLO burn.
3. Compare the failing case with a successful case before editing settings.
4. If this is the cache or propagation drift branch, Separate DNS, CDN, browser, and application cache, then purge only the stale layer.
5. Re-check with the same command and URL, then record the normal output.

Actions by cause

• For healthcheck, apply the fix only after reproducing the same condition and saving the before/after evidence for this exact code.
• Separate liveness, readiness, synthetic checks, and user traffic signals.
• Add service, route, status, and deployment labels to metrics.
• Tune alert thresholds using recent normal traffic windows.
• Delay readiness until dependencies are actually available.
• Create a short incident note with query links for the same signal.
• For the cache or propagation drift branch, Separate DNS, CDN, browser, and application cache, then purge only the stale layer.

Evidence links

• Docker troubleshooting official
• Docker Compose reference official

Verification metadata

• operator-draft
• official-reference-linked
• 2026-07-23

Update queue

• Review cadence
  weekly-source-review
• Next enrichment
  Add one official-source check and one real output example for Docker healthcheck.

Environment-specific checks

• Shared hosting, proxies, VPNs, or CDN layers can change image issue from runtime environment issue results.
• Do not trust only the Docker UI; compare command output.
• Japanese hosting panels may show completion before DNS or SSL fully propagates.
• Test from both office and external networks.

Prevent it next time

• Store normal examples for logs, exit code, volume, network, port bind.
• Add volume permissions, port conflicts, image tag, healthcheck to the release checklist.
• Keep recurring errors in the same note format.
• Split alerts by error rate, latency, certificates, disk, and permission changes.

Related notes

• Docker container exited 137
• Docker port already allocated fix note
• Docker docker.sock fix note
• Docker port already allocated fix note 2
• Docker docker.sock fix note 2
• Docker port already allocated fix note 3
• Docker docker.sock fix note 3
• Docker port already allocated fix note 4