Merge pull request #285 from runpod-workers/feat/0.17.1

feat: update vllm to 0.17.1

.github/workflows/release.yml

@@ -76,11 +76,23 @@ jobs:
 
       - name: Release Summary
         run: |
-          echo "🚀 Release completed!"
+          echo "Release completed!"
           echo "Version: ${{ env.RELEASE_VERSION }}"
           echo "Docker Image: ${{ env.DOCKERHUB_REPO }}/${{ env.DOCKERHUB_IMG }}:${{ env.RELEASE_VERSION }}"
-          if [[ "${{ github.event_name }}" == "workflow_dispatch" ]]; then
-            echo "Trigger: Manual workflow dispatch"
-          else
-            echo "Trigger: GitHub release (tag: ${{ github.ref_name }})"
-          fi
+
+      - name: Notify Slack
+        run: |
+          curl -sf -X POST "${{ secrets.SLACK_WEBHOOK_URL }}" \
+            -H "Content-Type: application/json" \
+            -d '{
+              "text": ":rocket: New :runpod-new-whiteonpurple: Runpod worker-vllm release: *${{ env.RELEASE_VERSION }}*",
+              "blocks": [
+                {
+                  "type": "section",
+                  "text": {
+                    "type": "mrkdwn",
+                    "text": ":banana-dance: *New Release — worker-vllm ${{ env.RELEASE_VERSION }}*\n*Docker:* `${{ env.DOCKERHUB_REPO }}/${{ env.DOCKERHUB_IMG }}:${{ env.RELEASE_VERSION }}`\n<https://github.com/${{ github.repository }}/releases/tag/${{ env.RELEASE_VERSION }}|View release on GitHub>"
+                  }
+                }
+              ]
+            }'

.github/workflows/slack-pr-issue-notify.yml

@@ -0,0 +1,41 @@
+name: Slack PR Notifications
+
+on:
+  pull_request:
+    types: [opened]
+  issues:
+    types: [opened]
+
+permissions:
+    contents: read
+   
+jobs:
+  notify:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Notify Slack - New PR
+        if: github.event_name == 'pull_request'
+        run: |
+          curl -sf -X POST "${{ secrets.SLACK_WEBHOOK_URL }}" \
+            -H "Content-Type: application/json" \
+            -d '{
+              "text": ":rocket: New PR in worker-vllm: *${{ github.event.pull_request.title }}*",
+              "blocks": [
+                {
+                  "type": "section",
+                  "text": {
+                    "type": "mrkdwn",
+                    "text": ":rocket: *New Pull Request — worker-vllm*\n*<${{ github.event.pull_request.html_url }}|${{ github.event.pull_request.title }}>*\nOpened by *${{ github.event.pull_request.user.login }}*"
+                  }
+                },
+                {
+                  "type": "context",
+                  "elements": [
+                    {
+                      "type": "mrkdwn",
+                      "text": "${{ github.event.pull_request.base.ref }} ← ${{ github.event.pull_request.head.ref }}"
+                    }
+                  ]
+                }
+              ]
+            }'

.github/workflows/slack-vllm-monitor.yml

@@ -0,0 +1,73 @@
+name: Monitor vLLM Releases
+
+on:
+  schedule:
+    - cron: '0 0 * * *' # Every day at midnight
+  workflow_dispatch:
+
+
+permissions:
+    contents: read
+   
+jobs:
+  check-vllm-release:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Restore last known vLLM tag
+        uses: actions/cache/restore@v4
+        with:
+          path: .vllm-last-tag
+          key: vllm-tag-${{ github.run_id }}
+          restore-keys: vllm-tag-
+
+      - name: Get latest vLLM release
+        id: vllm
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          response=$(curl -sf https://api.github.com/repos/vllm-project/vllm/releases/latest \
+            -H "Authorization: Bearer $GH_TOKEN")
+          echo "tag=$(echo "$response" | jq -r '.tag_name')" >> $GITHUB_OUTPUT
+          echo "url=$(echo "$response" | jq -r '.html_url')" >> $GITHUB_OUTPUT
+          echo "name=$(echo "$response" | jq -r '.name')" >> $GITHUB_OUTPUT
+
+      - name: Check if new release
+        id: check
+        run: |
+          last=$(cat .vllm-last-tag 2>/dev/null || echo "")
+          current="${{ steps.vllm.outputs.tag }}"
+          echo "Last: $last  Current: $current"
+          if [ -n "$current" ] && [ "$last" != "$current" ]; then
+            echo "is_new=true" >> $GITHUB_OUTPUT
+          else
+            echo "is_new=false" >> $GITHUB_OUTPUT
+          fi
+
+      - name: Notify Slack
+        if: steps.check.outputs.is_new == 'true'
+        run: |
+          curl -sf -X POST "${{ secrets.SLACK_WEBHOOK_URL }}" \
+            -H "Content-Type: application/json" \
+            -d '{
+              "text": ":rocket: New vLLM release: *${{ steps.vllm.outputs.tag }}*",
+              "blocks": [
+                {
+                  "type": "section",
+                  "text": {
+                    "type": "mrkdwn",
+                    "text": ":rocket: *New vLLM Release: ${{ steps.vllm.outputs.tag }}*\n<${{ steps.vllm.outputs.url }}|View on GitHub>"
+                  }
+                }
+              ]
+            }'
+
+      - name: Save new tag
+        if: steps.check.outputs.is_new == 'true'
+        run: echo "${{ steps.vllm.outputs.tag }}" > .vllm-last-tag
+
+      - name: Update cache
+        if: steps.check.outputs.is_new == 'true'
+        uses: actions/cache/save@v4
+        with:
+          path: .vllm-last-tag
+          key: vllm-tag-${{ steps.vllm.outputs.tag }}

Dockerfile

@@ -2,7 +2,7 @@ FROM nvidia/cuda:12.9.1-base-ubuntu22.04
 
 RUN apt-get update -y \
     && apt-get install -y python3-pip curl \
-    && curl -LsSf https://astral.sh/uv/0.10.9/install.sh  | sh
+    && curl -LsSf https://astral.sh/uv/install.sh  | sh
 
 ENV PATH="/root/.local/bin:$PATH"
 
@@ -10,7 +10,7 @@ RUN ldconfig /usr/local/cuda-12.9/compat/
 
 # Install vLLM with FlashInfer - use CUDA 12.9 PyTorch wheels
 RUN uv pip install --system "packaging>=24.2" && \
-    uv pip install --system "vllm[flashinfer]==0.16.0" --extra-index-url https://download.pytorch.org/whl/cu129
+    uv pip install --system "vllm[flashinfer]==0.17.1" --extra-index-url https://download.pytorch.org/whl/cu129
 
 # Install additional Python dependencies (after vLLM to avoid PyTorch version conflicts)
 COPY builder/requirements.txt /requirements.txt

builder/requirements.txt

@@ -3,7 +3,7 @@ pandas
 pyarrow
 runpod
 huggingface-hub
-lmcache==0.4.1
+lmcache==0.4.2
 packaging>=24.2
 typing-extensions>=4.8.0
 pydantic