LCORE-333: Lightspeed core needs to fully support RHEL AI LLM provider #714
Conversation
WalkthroughAdds RHEL AI (vLLM) support: new daily/on-demand GitHub Actions workflow for RHEL AI e2e tests, changes curl to fail-on-HTTP-error in an existing RHAIIS workflow, docker-compose env vars for RHEL AI, docs updates, and a new e2e config Changes
Sequence Diagram(s)sequenceDiagram
actor GitHub as GitHub Actions
participant Workflow as RHEL AI Workflow
participant Repo as Repository
participant Config as Config Generation
participant Endpoint as RHEL AI Endpoint
participant Docker as Docker Compose
participant Tests as E2E Tests
GitHub->>Workflow: Trigger (daily / workflow_dispatch)
Workflow->>Repo: Checkout code
Workflow->>Config: Generate/select `lightspeed-stack.yaml` / `run-rhelai.yaml`
Workflow->>Endpoint: Connectivity check (curl -f) %% explicit HTTP failure on non-2xx
alt endpoint reachable
Workflow->>Docker: docker-compose up (start services)
Docker->>Docker: wait / health checks
Workflow->>Endpoint: Quick local connectivity test
Workflow->>Tests: Run e2e tests (uv sync, make test-e2e)
Tests-->>Workflow: Test results
else endpoint/error
Workflow->>Workflow: Fail early and dump logs
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Areas to pay attention to:
Possibly related PRs
Suggested reviewers
Poem
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches🧪 Generate unit tests (beta)
📜 Recent review detailsConfiguration used: CodeRabbit UI Review profile: CHILL Plan: Pro 📒 Files selected for processing (2)
🚧 Files skipped from review as they are similar to previous changes (2)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (4)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (5)
.github/workflows/e2e_tests_rhaiis.yaml (1)
119-125: Port handling inconsistency with RHEL AI workflow.The RHAIIS connectivity test hardcodes port 8000, while the new RHEL AI workflow (line 126 in
.github/workflows/e2e_tests_rhelai.yaml) uses a parameterized${RHEL_AI_PORT}. For consistency and flexibility, consider parameterizing RHAIIS_PORT similarly.Additionally, both workflows should validate that secrets are set before attempting curl requests to provide clearer error messages if credentials are missing.
docker-compose.yaml (1)
20-22: Environment variables added correctly, but port handling differs from RHAIIS pattern.The RHEL_AI environment variables are properly added. However, note that RHAIIS omits
RHAIIS_PORT(hardcoded as 8000 in workflows), while RHEL_AI includesRHEL_AI_PORT. Consider whether RHAIIS should also have a parameterized port for consistency and future flexibility..github/workflows/e2e_tests_rhelai.yaml (2)
80-119: Run.yaml configuration step is solid, but sed patterns may be brittle.The script correctly selects and copies the environment-specific config file. However, the sed patterns (lines 114-115) assume specific path formats. If paths in configuration files vary, these patterns might fail silently or miss updates.
Consider making path normalization more robust (e.g., validate the result or use a dedicated tool).
120-127: Connectivity test properly uses parameterized port but lacks fallback messaging.The curl connectivity test correctly uses
-fflag and parameterized${RHEL_AI_PORT}, improving on the RHAIIS workflow. However, if the external endpoint is unreachable, the entire job fails without clear messaging about which secret/endpoint was unavailable.Consider adding explicit error checking with descriptive output.
Apply a diff to improve error messaging:
- name: Test RHEL_AI connectivity env: RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} run: | + if [ -z "$RHEL_AI_URL" ] || [ -z "$RHEL_AI_PORT" ] || [ -z "$RHEL_AI_API_KEY" ]; then + echo "❌ Missing required RHEL_AI secrets (RHEL_AI_URL, RHEL_AI_PORT, RHEL_AI_API_KEY)" + exit 1 + fi curl -f ${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models -H "Authorization: Bearer ${RHEL_AI_API_KEY}"tests/e2e/configs/run-rhelai.yaml (1)
75-81: RHEL AI vLLM provider configuration looks structurally sound, but review security and token settings.The provider configuration correctly uses environment variable substitution (
${env.RHEL_AI_URL}, etc.). A few observations:
- Line 80:
tls_verify: falsedisables TLS certificate verification—acceptable for E2E tests against internal endpoints, but ensure this isn't propagated to production configurations.- Line 81:
max_tokens: 2048is hardcoded. Verify this aligns with the model's capabilities or consider making it configurable for future flexibility.Apply a diff to document the security consideration:
- provider_id: vllm provider_type: remote::vllm config: url: http://${env.RHEL_AI_URL}:${env.RHEL_AI_PORT}/v1/ api_token: ${env.RHEL_AI_API_KEY} - tls_verify: false + tls_verify: false # E2E testing only; use true in production max_tokens: 2048
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
.github/workflows/e2e_tests_rhaiis.yaml(1 hunks).github/workflows/e2e_tests_rhelai.yaml(1 hunks)README.md(1 hunks)docker-compose.yaml(1 hunks)docs/providers.md(1 hunks)tests/e2e/configs/run-rhelai.yaml(1 hunks)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)
- GitHub Check: build-pr
- GitHub Check: e2e_tests (ci)
- GitHub Check: e2e_tests (azure)
🔇 Additional comments (5)
docs/providers.md (1)
62-65: Provider documentation is consistent and complete.The new RHEL AI (vLLM) entry is properly formatted and aligns with the version tested (1.5.2) mentioned in the PR description. The dependency on
openaiis correct for the OpenAI-compatible vLLM API.README.md (1)
123-129: LLM Compatibility table entry is consistent and correctly linked.The new RHEL AI (vLLM) entry properly references the test configuration at
tests/e2e/configs/run-rhelai.yamland maintains consistency with the existing RHAIIS entry and documentation indocs/providers.md.tests/e2e/configs/run-rhelai.yaml (2)
130-140: Model definitions align with documentation and are properly wired to providers.The embedding and LLM models correctly reference their respective providers (sentence-transformers and vllm). The model IDs and provider model IDs match documentation and support the test scenario.
1-27: Configuration metadata and API list are comprehensive.The file properly declares all required APIs for RHEL AI E2E testing. Path normalization in the workflow handles the relative paths correctly.
.github/workflows/e2e_tests_rhelai.yaml (1)
160-167: Review comment is incorrect.The concern assumes
run.yamlmust be loaded by the test-e2e target, but this misunderstands the architecture. GitHub Actions does NOT automatically "source" a run.yaml from your repository root, and in this case, run.yaml doesn't need to be. The workflow correctly:
- Copies
run.yamlfromtests/e2e/configs/run-$ENVIRONMENT.yamlto be used by docker-compose (for llama-stack service configuration)- E2E tests connect to services via environment variables (
E2E_LSC_HOSTNAME,E2E_LSC_PORT), not by loading run.yamlThe
make test-e2etarget uses behave and environment variables—no explicit run.yaml reference is needed or expected.Likely an incorrect or invalid review comment.
are-ces
force-pushed
the
rhelai-support
branch
from
52d2615 to
f90ef53
Compare
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
.github/workflows/e2e_tests_rhelai.yaml (2)
146-159: Consider replacing fixed sleep with health checks or retry logic.The 20-second fixed sleep (line 149) is a blunt instrument. Services may take longer under load or in CI runners, causing flaky tests. A better approach would use health checks or retry logic with exponential backoff.
Optional improvement (if flakiness is observed):
- name: Wait for services run: | echo "Waiting for services to be healthy..." - sleep 20 # adjust depending on boot time + max_retries=30 + retry_count=0 + while [ $retry_count -lt $max_retries ]; do + if curl -f http://localhost:8080/v1/models >/dev/null 2>&1 && \ + curl -f http://llama-stack:8321/v1/models >/dev/null 2>&1; then + echo "✓ Services are healthy" + exit 0 + fi + retry_count=$((retry_count + 1)) + echo "Attempt $retry_count/$max_retries: services not ready, waiting..." + sleep 2 + done + echo "✗ Services failed to become healthy after $((max_retries * 2))s" + exit 1This avoids unnecessary delays when services are ready quickly and provides faster feedback when they're not.
16-19: Consolidate redundant environment variable declarations.The RHEL AI secrets are declared in three separate places (job-level env, test connectivity step env, and run service step env). While functional, this violates DRY and creates maintenance risk. Consider declaring once at the job level since they're used consistently.
env: RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} steps: # ... checkout steps ... - name: Test RHEL_AI connectivity - env: - RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} - RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} - RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} run: | # ... uses inherited env vars ... - name: Run service manually - env: - RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} - RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} - RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} run: | # ... uses inherited env vars ...Also applies to: 121-124, 130-132
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
.github/workflows/e2e_tests_rhaiis.yaml(1 hunks).github/workflows/e2e_tests_rhelai.yaml(1 hunks)README.md(1 hunks)docker-compose.yaml(1 hunks)docs/providers.md(1 hunks)tests/e2e/configs/run-rhelai.yaml(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (4)
- .github/workflows/e2e_tests_rhaiis.yaml
- docs/providers.md
- README.md
- docker-compose.yaml
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)
- GitHub Check: build-pr
- GitHub Check: e2e_tests (azure)
- GitHub Check: e2e_tests (ci)
🔇 Additional comments (8)
.github/workflows/e2e_tests_rhelai.yaml (4)
1-33: Workflow structure and trigger configuration look solid.The daily schedule, on-demand dispatch option, and PR-aware checkout with proper credential handling follow GitHub Actions security best practices.
48-79: Lightspeed-stack service configuration is appropriate for E2E testing.The remote llama-stack client setup and test authentication are suitable for this use case.
160-178: Test execution and failure diagnostics are well-structured.The test invocation and conditional logging on failure are appropriate. The dual logging of both llama-stack and lightspeed-stack containers will aid in debugging.
113-115: Review comment is incorrect and should be disregarded.The sed patterns are appropriate for the actual configuration file. Analysis of
run-rhelai.yamlshows:
- The first sed pattern
db_path: \.llama/distributionscorrectly matches all 8 existing db_path lines that use this relative path- The second sed pattern
db_path: tmp/does not match any lines in the config—the only "tmp/" in the file isstorage_dir: /tmp/llama-stack-files, which doesn't match the pattern. The second sed is a harmless no-op- The pattern includes the key name "db_path:" which provides adequate specificity to prevent false matches in different contexts
The suggested fixes in the review would actually break the intended replacements by searching for non-existent patterns like
db_path: ../llama/distributions.Likely an incorrect or invalid review comment.
tests/e2e/configs/run-rhelai.yaml (4)
1-16: API configuration is appropriate for comprehensive E2E testing.The selected APIs cover inference, evaluation, safety, and operational concerns needed for meaningful RHEL AI integration testing.
29-117: Provider configuration is comprehensive and appropriate.The mix of inline (local) and remote providers with proper configuration for each use case (agents, datasets, evaluation, safety, scoring, telemetry) provides good E2E test coverage. The Braintrust placeholder credentials are acceptable for test configuration.
118-140: Server, shields, and model configuration are well-structured.The minimal server setup, Llama-guard shield integration, and selection of sentence-transformers for embeddings plus Llama-3.1-8B for inference via vLLM provide solid test coverage for the RHEL AI integration.
1-140: No issues found. The environment variable syntax${env.VARIABLE_NAME}is the standard pattern used consistently across all llama-stack configuration files in the codebase (run-azure.yaml, run-ci.yaml, run-rhaiis.yaml, and run-rhelai.yaml). This confirms the syntax is fully supported and compatible with llama-stack's YAML parser.
| - name: Test RHEL_AI connectivity | ||
| env: | ||
| RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} | ||
| RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} | ||
| RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} | ||
| run: | | ||
| curl -f ${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models -H "Authorization: Bearer ${RHEL_AI_API_KEY}" | ||
|
|
There was a problem hiding this comment.
Add quoting and error handling to curl command.
Environment variables used in the curl command (line 126) are unquoted and lack null-value checks. If RHEL_AI_URL or RHEL_AI_PORT are empty (e.g., secrets not configured), the command will silently construct an invalid URL. The curl output also lacks error reporting.
Apply this diff:
- name: Test RHEL_AI connectivity
env:
RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }}
RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }}
RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }}
run: |
- curl -f ${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models -H "Authorization: Bearer ${RHEL_AI_API_KEY}"
+ if [ -z "$RHEL_AI_URL" ] || [ -z "$RHEL_AI_PORT" ] || [ -z "$RHEL_AI_API_KEY" ]; then
+ echo "Error: RHEL_AI_URL, RHEL_AI_PORT, or RHEL_AI_API_KEY not set"
+ exit 1
+ fi
+ curl -f "${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models" \
+ -H "Authorization: Bearer ${RHEL_AI_API_KEY}" || {
+ echo "Connectivity test failed - RHEL AI endpoint unreachable"
+ exit 1
+ }📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| - name: Test RHEL_AI connectivity | |
| env: | |
| RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} | |
| RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} | |
| RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} | |
| run: | | |
| curl -f ${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models -H "Authorization: Bearer ${RHEL_AI_API_KEY}" | |
| - name: Test RHEL_AI connectivity | |
| env: | |
| RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} | |
| RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} | |
| RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} | |
| run: | | |
| if [ -z "$RHEL_AI_URL" ] || [ -z "$RHEL_AI_PORT" ] || [ -z "$RHEL_AI_API_KEY" ]; then | |
| echo "Error: RHEL_AI_URL, RHEL_AI_PORT, or RHEL_AI_API_KEY not set" | |
| exit 1 | |
| fi | |
| curl -f "${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models" \ | |
| -H "Authorization: Bearer ${RHEL_AI_API_KEY}" || { | |
| echo "Connectivity test failed - RHEL AI endpoint unreachable" | |
| exit 1 | |
| } |
🤖 Prompt for AI Agents
In .github/workflows/e2e_tests_rhelai.yaml around lines 120 to 127, the curl
command uses unquoted environment variables and does not check for empty values
or surface curl errors; update the workflow to validate that RHEL_AI_URL and
RHEL_AI_PORT (and RHEL_AI_API_KEY) are set before running curl, quote the
variables when constructing the URL and header, and run curl with -fS --retry
and capture/print HTTP status or fail explicitly so the job errors on missing
secrets or connection failures.
00c28d2 to
7391953
Compare
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (1)
.github/workflows/e2e_tests_rhelai.yaml (1)
122-125: Add quoting and validation to curl command (duplicate of previous review).The curl command uses unquoted environment variables that may be empty or unset. If
RHEL_AI_URL,RHEL_AI_PORT, orRHEL_AI_API_KEYare not configured as secrets, the command will fail silently or with a cryptic error message. Add explicit validation and proper quoting.- name: Test RHEL_AI connectivity run: | + if [ -z "$RHEL_AI_URL" ] || [ -z "$RHEL_AI_PORT" ] || [ -z "$RHEL_AI_API_KEY" ]; then + echo "Error: RHEL_AI_URL, RHEL_AI_PORT, or RHEL_AI_API_KEY not set" + exit 1 + fi - curl -f ${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models -H "Authorization: Bearer ${RHEL_AI_API_KEY}" + curl -f "${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models" \ + -H "Authorization: Bearer ${RHEL_AI_API_KEY}" || { + echo "Connectivity test failed - RHEL AI endpoint unreachable" + exit 1 + }
🧹 Nitpick comments (1)
.github/workflows/e2e_tests_rhelai.yaml (1)
140-143: Consider making the service wait time configurable.The hardcoded 20-second sleep is reasonable for typical startup, but services (llama-stack, lightspeed-stack) may have variable startup times depending on model loading and system resources. Consider adding a health check loop or allowing this to be overridden via an environment variable or job parameter for better resilience.
- name: Wait for services run: | echo "Waiting for services to be healthy..." - sleep 20 # adjust depending on boot time + WAIT_TIME=${SERVICE_WAIT_TIME:-20} + echo "Waiting ${WAIT_TIME} seconds for services to stabilize..." + sleep "$WAIT_TIME"
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
.github/workflows/e2e_tests_rhelai.yaml(1 hunks)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (2)
- GitHub Check: e2e_tests (ci)
- GitHub Check: e2e_tests (azure)
are-ces
force-pushed
the
rhelai-support
branch
from
cf442c6 to
5a66916
Compare
There was a problem hiding this comment.
Actionable comments posted: 0
♻️ Duplicate comments (2)
tests/e2e/configs/run-rhelai.yaml (1)
75-81: TLS verification should be configurable via environment variable.Line 80 hardcodes
tls_verify: false, which disables TLS certificate validation. While acceptable for development/test environments, this should be configurable to support different security postures.Following the pattern used elsewhere in the codebase (e.g.,
examples/vertexai-run.yaml), make this environment-driven:- provider_id: vllm provider_type: remote::vllm config: url: http://${env.RHEL_AI_URL}:${env.RHEL_AI_PORT}/v1/ api_token: ${env.RHEL_AI_API_KEY} - tls_verify: false + tls_verify: ${env.RHEL_AI_TLS_VERIFY:=false} max_tokens: 2048Then add
RHEL_AI_TLS_VERIFYto.github/workflows/e2e_tests_rhelai.yaml(env section) anddocker-compose.yamlso it can be controlled at runtime..github/workflows/e2e_tests_rhelai.yaml (1)
121-125: Add null checks and quote variables in connectivity test.The curl command lacks validation that required secrets are set, and uses unquoted variables which could be problematic if they contain special characters.
- name: Test RHEL_AI connectivity + env: + RHEL_AI_URL: ${{ secrets.RHEL_AI_URL }} + RHEL_AI_PORT: ${{ secrets.RHEL_AI_PORT }} + RHEL_AI_API_KEY: ${{ secrets.RHEL_AI_API_KEY }} run: | - echo $RHEL_AI_MODEL - curl -f ${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models -H "Authorization: Bearer ${RHEL_AI_API_KEY}" + if [ -z "$RHEL_AI_URL" ] || [ -z "$RHEL_AI_PORT" ] || [ -z "$RHEL_AI_API_KEY" ]; then + echo "Error: RHEL_AI_URL, RHEL_AI_PORT, or RHEL_AI_API_KEY not set" + exit 1 + fi + curl -f "${RHEL_AI_URL}:${RHEL_AI_PORT}/v1/models" \ + -H "Authorization: Bearer ${RHEL_AI_API_KEY}" || { + echo "Connectivity test failed - RHEL AI endpoint unreachable" + exit 1 + }
🧹 Nitpick comments (1)
.github/workflows/e2e_tests_rhaiis.yaml (1)
124-124: Quote environment variables in curl command for robustness.While the
-fflag correctly fails on HTTP errors, unquoted variables should be quoted to prevent shell expansion issues if the values contain special characters or spaces.- curl -f ${RHAIIS_URL}:${RHAIIS_PORT}/v1/models -H "Authorization: Bearer ${RHAIIS_API_KEY}" + curl -f "${RHAIIS_URL}:${RHAIIS_PORT}/v1/models" \ + -H "Authorization: Bearer ${RHAIIS_API_KEY}"
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
.github/workflows/e2e_tests_rhaiis.yaml(1 hunks).github/workflows/e2e_tests_rhelai.yaml(1 hunks)README.md(1 hunks)docker-compose.yaml(1 hunks)docs/providers.md(1 hunks)tests/e2e/configs/run-rhelai.yaml(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- README.md
- docs/providers.md
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (2)
- GitHub Check: e2e_tests (azure)
- GitHub Check: e2e_tests (ci)
🔇 Additional comments (4)
docker-compose.yaml (1)
20-23: Environment variables for RHEL AI added cleanly.The addition of RHEL_AI_URL, RHEL_AI_PORT, RHEL_AI_API_KEY, and RHEL_AI_MODEL follows the existing pattern and naming conventions. Good alignment with the new RHEL AI provider integration.
.github/workflows/e2e_tests_rhelai.yaml (3)
81-119: Configuration setup logic is sound.The configuration selection and path transformation steps include good defensive checks (directory exists, file found, readable alternatives on failure) and clear diagnostics.
126-152: Service startup and health check sequence is well-structured.The workflow includes appropriate checks (services started without exit codes, wait period, connectivity test) with good error diagnostics (docker logs on failure). This progressive verification approach is sound.
154-175: Test execution and failure logging are well-organized.The workflow installs dependencies efficiently (uv sync), runs the standard test target, and provides targeted service logs on failure. The conditional logging step (if: failure()) is appropriate and keeps logs focused.
3 participants
Description
Added e2e tests for RHEL AI (1.5.2), updated docs.
e2e tests are set to run every day at midnight UTC.
Type of change
Related Tickets & Documents
Checklist before requesting a review
Testing
E2E test for RHEL AI shows that the inference works, thus is supported.
Summary by CodeRabbit
New Features
Documentation
Chores