Make link check consistent (open-telemetry#1986)

Author: breedx-splk

.github/scripts/dependencies.dockerfile

@@ -0,0 +1,3 @@
+# this file exists so that Renovate can auto-update docker image versions that are then used elsewhere
+
+FROM lycheeverse/lychee:sha-2aa22f8@sha256:2e3786630482c41f9f2dd081e06d7da1c36d66996e8cf6573409b8bc418d48c4 AS lychee

.github/scripts/link-check.sh

@@ -0,0 +1,35 @@
+#!/bin/bash
+
+set -e
+
+export MSYS_NO_PATHCONV=1 # for Git Bash on Windows
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+LYCHEE_CONFIG="$SCRIPT_DIR/../../.lychee.toml"
+DEPENDENCIES_DOCKERFILE="$SCRIPT_DIR/dependencies.dockerfile"
+
+# Extract lychee version from dependencies.dockerfile
+LYCHEE_VERSION=$(grep "FROM lycheeverse/lychee:" "$DEPENDENCIES_DOCKERFILE" | sed 's/.*FROM lycheeverse\/lychee:\([^ ]*\).*/\1/')
+
+# Build the lychee command with optional GitHub token
+CMD="lycheeverse/lychee:$LYCHEE_VERSION --verbose --config $(basename "$LYCHEE_CONFIG")"
+
+# Add GitHub token if available
+if [[ -n "$GITHUB_TOKEN" ]]; then
+    CMD="$CMD --github-token $GITHUB_TOKEN"
+fi
+
+# Add the target directory
+CMD="$CMD ."
+
+# Determine if we should allocate a TTY
+DOCKER_FLAGS="--rm --init"
+if [[ -t 0 ]]; then
+    DOCKER_FLAGS="$DOCKER_FLAGS -it"
+else
+    DOCKER_FLAGS="$DOCKER_FLAGS -i"
+fi
+
+# Run lychee with proper signal handling
+# shellcheck disable=SC2086
+exec docker run $DOCKER_FLAGS -v "$(dirname "$LYCHEE_CONFIG")":/data -w /data $CMD

.github/workflows/build.yml

@@ -118,8 +118,12 @@ jobs:
           name: integration-test-results
           path: jmx-metrics/build/reports/tests/integrationTest
 
-  markdown-link-check:
-    uses: ./.github/workflows/reusable-markdown-link-check.yml
+  link-check:
+    # release branches are excluded to avoid unnecessary maintenance if external links break
+    # (and also because the README.md might need update on release branches before the release
+    # download has been published)
+    if: "!startsWith(github.ref_name, 'release/')"
+    uses: ./.github/workflows/reusable-link-check.yml
 
   markdown-lint-check:
     uses: ./.github/workflows/reusable-markdown-lint.yml
@@ -142,7 +146,7 @@ jobs:
     # and so would not short-circuit if used in the second-last position
     name: publish-snapshots${{ (github.ref_name != 'main' || github.repository != 'open-telemetry/opentelemetry-java-contrib') && ' (skipped)' || '' }}
     needs:
-      # intentionally not blocking snapshot publishing on markdown-link-check or misspell-check
+      # intentionally not blocking snapshot publishing on link-check or misspell-check
       - build
       - integration-test
     runs-on: ubuntu-latest

.github/workflows/reusable-link-check.yml

@@ -0,0 +1,18 @@
+name: Reusable - Link check
+
+on:
+  workflow_call:
+
+permissions:
+  contents: read
+
+jobs:
+  link-check:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+
+      - name: Link check
+        env:
+          GITHUB_TOKEN: ${{ github.token }}
+        run: ./.github/scripts/link-check.sh

.github/workflows/reusable-markdown-link-check.yml

@@ -0,0 +1,16 @@
+timeout = 30
+retry_wait_time = 5
+max_retries = 6
+max_concurrency = 4
+
+# Check link anchors
+include_fragments = true
+
+# excluding links to pull requests and issues is done for performance
+# sonatype snapshots are currrently unbrowseable
+exclude = [
+  "^https://github.com/open-telemetry/opentelemetry-java-contrib/(issues|pull)/\\d+$",
+  '^https://central.sonatype.com/service/rest/repository/browse/maven-snapshots/io/opentelemetry/contrib/$',
+]
+
+