Add daily Qdrant keepalive workflow

[automationpi]

Pings Qdrant cluster daily at 8am UTC to prevent free tier suspension. Also supports manual trigger.

[github-actions]

mergelore findings

Reviewed 0 past PRs · 0 relevant decisions found · confidence threshold 0.5

⚠️ Potential conflicts

Hardcoded HTTP 200 status check may miss valid Qdrant responses · confidence: 0.60

The keepalive workflow checks if the Qdrant API responds with exactly HTTP 200. However, other 2xx status codes (201, 202, etc.) or even 3xx redirects could indicate a healthy cluster. The current logic treats anything other than 200 as a warning, which could produce false positives and noise in CI/CD monitoring. A more robust check would validate the response is in the 2xx range.

/mergelore-acknowledge · /mergelore-override [reason] · /mergelore-update-record

Missing error handling for curl command failures · confidence: 0.65

The curl command itself could fail (network timeout, DNS resolution failure, etc.) and still exit with a non-zero status, but the script doesn't distinguish between curl errors and non-200 HTTP responses. If curl fails completely, $STATUS may be empty or set to a curl error code, which won't match '200' but also won't clearly indicate what went wrong. Consider adding explicit error handling for curl execution failures.

/mergelore-acknowledge · /mergelore-override [reason] · /mergelore-update-record

---

ℹ️ Related context

• Secret exposure risk in workflow logs (confidence: 0.50): The workflow references QDRANT_URL and QDRANT_API_KEY as GitHub secrets, which is correct. However, if the curl command or echo statements were to fail or be logged, sensitive information could be exposed. While the current implementation appears safe (no direct logging of secrets), this is a structural pattern to monitor given that secrets are used in shell commands.

---

mergelore v0.1.0 · docs