fix(ci): replace curl health-check with bash TCP check for qdrant container

[vineethwilson15]

Problem

The integration tests job was failing with:

Service container qdrant failed.
Error: Failed to initialize container qdrant/qdrant:latest

Qdrant itself started successfully (HTTP on 6333, gRPC on 6334), but the
GitHub Actions service container health check was reporting failure.

Root cause: qdrant/qdrant:latest v1.17.0+ uses debian:bookworm-slim
as its base image, which does not ship curl. The configured health
check command curl -sf http://localhost:6333/healthz exited with
command not found, causing GitHub Actions to mark the container as
unhealthy and abort the job.

Fix

Replace with a pure bash TCP connection check that requires no external
tooling:

bash -c ':> /dev/tcp/localhost/6333'

This opens a TCP socket to port 6333. It succeeds as soon as Qdrant's HTTP
server is accepting connections — no curl, wget, or any other binary
required.

Testing

• Lint & unit test jobs are unaffected (no Docker services used)
• Add the run-integration label to this PR to trigger the integration
  tests job and verify the fix

Summary by CodeRabbit

• Chores
  • Improved service health check validation in CI/CD pipeline by updating the health probe method for more reliable service readiness detection.

[coderabbitai]

Walkthrough

The qdrant service health check in the CI workflow was modified from an HTTP-based curl probe to a TCP connection probe using bash. This changes how the service readiness is validated during testing.

Changes

Cohort / File(s)	Summary
CI Health Check .github/workflows/ci.yml	Modified qdrant service health check from curl-based HTTP probe to TCP-connect probe using bash, altering service validation method.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: replacing a curl-based health check with a bash TCP check for the qdrant container in CI workflows.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/qdrant-healthcheck-no-curl

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

.github/workflows/ci.yml (1)

105-110: ⚠️ Potential issue | 🟠 Major

Pin Qdrant to a tested image tag to maintain dual-store consistency.

Neo4j is pinned to 5.18, but Qdrant uses latest. The health check depends on Bash-specific /dev/tcp, which can break on future base image changes. Pin Qdrant to the version validated by this PR to keep Neo4j and Qdrant versioning in sync.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In @.github/workflows/ci.yml around lines 105 - 110, Replace the qdrant service
image from the floating tag to the specific tested tag and change the
bash-specific health check: update the qdrant "image" value (the qdrant
service's image field) from "qdrant/qdrant:latest" to the pinned tag
used/validated by this PR, and replace the "options" health-cmd (which currently
uses bash /dev/tcp) with a more portable command (e.g., curl or wget against
localhost:6333/health) so the health check won't break if the base image drops
bash.

🧹 Nitpick comments (1)

.github/workflows/ci.yml (1)

109-110: Keep the readiness check at HTTP level.

A bare TCP connect is weaker than the old /healthz probe: it can go green on an open socket before the HTTP health endpoint is actually serving. You can keep the "no curl" fix and still probe /healthz over Bash TCP so the dual-store integration job waits for Qdrant itself, not just port 6333. (docs.docker.com)

♻️ Proposed fix

-          --health-cmd "bash -c ':> /dev/tcp/localhost/6333'"
+          --health-cmd "bash -ec 'exec 3<>/dev/tcp/127.0.0.1/6333 && printf \"GET /healthz HTTP/1.1\r\nHost: 127.0.0.1\r\nConnection: close\r\n\r\n\" >&3 && read -r status <&3 && [[ $status == *\" 200 \"* ]]'"

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In @.github/workflows/ci.yml around lines 109 - 110, The current readiness check
uses a raw TCP connect in the workflow options (the options: >- block with
--health-cmd "bash -c ':> /dev/tcp/localhost/6333'"), which can succeed before
Qdrant's HTTP health endpoint is ready; change the --health-cmd to perform an
HTTP-level probe against /healthz over the same TCP socket (send an HTTP GET for
/healthz on port 6333 and verify an HTTP 200/1.1 response) so the job waits for
the actual HTTP health endpoint rather than just an open port.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Outside diff comments:
In @.github/workflows/ci.yml:
- Around line 105-110: Replace the qdrant service image from the floating tag to
the specific tested tag and change the bash-specific health check: update the
qdrant "image" value (the qdrant service's image field) from
"qdrant/qdrant:latest" to the pinned tag used/validated by this PR, and replace
the "options" health-cmd (which currently uses bash /dev/tcp) with a more
portable command (e.g., curl or wget against localhost:6333/health) so the
health check won't break if the base image drops bash.

---

Nitpick comments:
In @.github/workflows/ci.yml:
- Around line 109-110: The current readiness check uses a raw TCP connect in the
workflow options (the options: >- block with --health-cmd "bash -c ':>
/dev/tcp/localhost/6333'"), which can succeed before Qdrant's HTTP health
endpoint is ready; change the --health-cmd to perform an HTTP-level probe
against /healthz over the same TCP socket (send an HTTP GET for /healthz on port
6333 and verify an HTTP 200/1.1 response) so the job waits for the actual HTTP
health endpoint rather than just an open port.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 05b033fc-d304-44cf-821a-d3ba0d653815

📥 Commits

Reviewing files that changed from the base of the PR and between 303d07e and 663fa90.

📒 Files selected for processing (1)

• .github/workflows/ci.yml