ci: report image size changes on image PRs

.github/scripts/format-image-size-report.ts

@@ -0,0 +1,123 @@
+// Renders one workflow's <details> section of the shared "image size" PR comment.
+// The section is upserted into a single sticky comment by upsert-image-size-comment.js,
+// so each image workflow only ever touches its own foldable block.
+//
+// Usage: node format-image-size-report.ts <reports-dir> <section-title> [head-sha]
+// Runs natively on Node.js >=24 via built-in TypeScript type stripping (no build step).
+//
+// <reports-dir> contains one JSON file per built image (uploaded as artifacts by
+// the build matrix). Each file has the shape described by SizeReport below;
+// `currentBytes` is empty when there is no published baseline image to compare against.
+
+import { readdirSync, readFileSync } from 'node:fs';
+import { join } from 'node:path';
+
+interface SizeReport {
+    matrix?: Record<string, string>;
+    baseImage?: string;
+    // Extra variant label not derivable from the matrix (e.g. "slim").
+    variant?: string;
+    currentBytes?: string;
+    newBytes?: string;
+}
+
+// Matrix keys we never want to surface as a "variant" in the table.
+const IGNORED_VERSION_KEYS = new Set<string>([
+    'node-version',
+    'python-version',
+    'apify-version',
+    'crawlee-version',
+    'latest-node-version',
+    'latest-python-version',
+]);
+
+const MiB = 1024 * 1024;
+
+function readReports(reportsDir: string): SizeReport[] {
+    let files: string[];
+    try {
+        files = readdirSync(reportsDir).filter((file) => file.endsWith('.json'));
+    } catch {
+        return [];
+    }
+
+    const reports: SizeReport[] = [];
+    for (const file of files) {
+        try {
+            reports.push(JSON.parse(readFileSync(join(reportsDir, file), 'utf8')) as SizeReport);
+        } catch {
+            // Skip unreadable/corrupt report files rather than failing the whole comment.
+        }
+    }
+    return reports;
+}
+
+function formatSize(bytes: string | undefined): string | null {
+    const n = Number(bytes);
+    if (!bytes || !Number.isFinite(n)) return null;
+    return `${(n / MiB).toFixed(1)} MiB`;
+}
+
+function formatDelta(currentBytes: string | undefined, newBytes: string | undefined): string {
+    const current = Number(currentBytes);
+    const next = Number(newBytes);
+    if (!currentBytes || !Number.isFinite(current) || !Number.isFinite(next)) {
+        return '🆕 _new image_';
+    }
+    const diff = next - current;
+    const pct = current === 0 ? 0 : (diff / current) * 100;
+    const sign = diff > 0 ? '+' : diff < 0 ? '−' : '';
+    const emoji = diff > 0 ? '🔺' : diff < 0 ? '🔻' : '➖';
+    const absMiB = (Math.abs(diff) / MiB).toFixed(1);
+    return `${emoji} ${sign}${absMiB} MiB (${sign}${Math.abs(pct).toFixed(1)}%)`;
+}
+
+function describeVariant(matrix: Record<string, string> = {}): string {
+    const variants: string[] = [];
+    for (const [key, value] of Object.entries(matrix)) {
+        if (!value || IGNORED_VERSION_KEYS.has(key) || !key.endsWith('-version')) continue;
+        variants.push(`${key.replace(/-version$/, '')} ${value}`);
+    }
+    return variants.join(', ');
+}
+
+function render(reports: SizeReport[], title: string, headSha: string): string {
+    const lines: string[] = [];
+    const measuredAt = headSha ? ` <sub>(at ${headSha.slice(0, 7)})</sub>` : '';
+
+    if (reports.length === 0) {
+        lines.push('<details>');
+        lines.push(`<summary><b>${title}</b> — no image size data collected${measuredAt}</summary>`);
+        lines.push('');
+        lines.push('No images were built in this run.');
+        lines.push('</details>');
+        lines.push('');
+        return lines.join('\n');
+    }
+
+    lines.push('<details>');
+    lines.push(`<summary><b>${title}</b> — ${reports.length} image${reports.length === 1 ? '' : 's'}${measuredAt}</summary>`);
+    lines.push('');
+    lines.push('| Image | Variant | Current | New | Δ |');
+    lines.push('| --- | --- | ---: | ---: | --- |');
+
+    const rows = reports
+        .map((report) => ({
+            image: report.baseImage || '(unknown)',
+            variant: [describeVariant(report.matrix), report.variant].filter(Boolean).join(', '),
+            current: formatSize(report.currentBytes) ?? '_n/a_',
+            next: formatSize(report.newBytes) ?? '_n/a_',
+            delta: formatDelta(report.currentBytes, report.newBytes),
+        }))
+        .sort((a, b) => a.image.localeCompare(b.image) || a.variant.localeCompare(b.variant));
+
+    for (const row of rows) {
+        lines.push(`| \`${row.image}\` | ${row.variant || '—'} | ${row.current} | ${row.next} | ${row.delta} |`);
+    }
+
+    lines.push('</details>');
+    lines.push('');
+    return lines.join('\n');
+}
+
+process.stdout.write(render(readReports(process.argv[2] ?? ''), process.argv[3] ?? 'Image sizes', process.argv[4] ?? ''));

.github/scripts/upsert-image-size-comment.js

@@ -0,0 +1,48 @@
+// Upserts one workflow's section into the single shared "image size" PR comment.
+//
+// Every image workflow renders its own <details> block (format-image-size-report.ts) and
+// calls this from actions/github-script. The comment is identified by MARKER; each section
+// is delimited by per-workflow markers so a workflow only ever rewrites its own block.
+// Concurrent writers are serialized by the size-report jobs sharing a repo-wide
+// `concurrency` group, so the read-modify-write below cannot lose sections.
+
+const MARKER = '<!-- image-size-report -->';
+const PREAMBLE = `${MARKER}
+### 📦 Image size report
+
+Built images compared against the currently published rolling tag for the same runtime \
+version (e.g. \`apify/actor-node:22\`; \`-slim\` variants against the \`-slim\` tag). Sizes \
+are the **uncompressed** on-disk size reported by \`docker image inspect\`, so they will be \
+larger than the compressed download size shown on Docker Hub. Only workflows triggered by \
+this PR's changes report a section.
+`;
+
+module.exports = async ({ github, context }, sectionKey, sectionBody) => {
+    if (!/^[a-z][a-z0-9-]*$/.test(sectionKey)) throw new Error(`Invalid section key: ${sectionKey}`);
+
+    const { owner, repo } = context.repo;
+    const issue_number = context.issue.number;
+    const begin = `<!-- section:${sectionKey} -->`;
+    const end = `<!-- /section:${sectionKey} -->`;
+    const section = `${begin}\n${sectionBody.trim()}\n${end}`;
+
+    const comments = await github.paginate(github.rest.issues.listComments, {
+        owner,
+        repo,
+        issue_number,
+        per_page: 100,
+    });
+    const existing = comments.find((comment) => comment.body?.includes(MARKER));
+
+    if (!existing) {
+        await github.rest.issues.createComment({ owner, repo, issue_number, body: `${PREAMBLE}\n${section}\n` });
+        return;
+    }
+
+    const sectionPattern = new RegExp(`${begin}[\\s\\S]*?${end}`);
+    const body = sectionPattern.test(existing.body)
+        ? existing.body.replace(sectionPattern, section)
+        : `${existing.body.trimEnd()}\n\n${section}\n`;
+
+    await github.rest.issues.updateComment({ owner, repo, comment_id: existing.id, body });
+};

.github/workflows/release-node-playwright.yaml

@@ -278,6 +278,50 @@ jobs:
       - name: Test slim image
         run: docker run ${{ fromJson(steps.prepare-slim-tags.outputs.result).firstImageName }}
 
+      - name: Measure image size
+        id: image-size
+        if: github.event_name == 'pull_request'
+        env:
+          MATRIX_JSON: ${{ toJSON(matrix) }}
+          NEW_IMAGE: ${{ fromJson(steps.prepare-tags.outputs.result).firstImageName }}
+          BASE_IMAGE: apify/actor-${{ matrix.image-name }}:${{ matrix.node-version }}
+          NEW_SLIM_IMAGE: ${{ fromJson(steps.prepare-slim-tags.outputs.result).firstImageName }}
+          BASE_SLIM_IMAGE: apify/actor-${{ matrix.image-name }}:${{ matrix.node-version }}-slim
+        run: |
+          set -euo pipefail
+          mkdir -p size-report
+          id="$(printf '%s' "$MATRIX_JSON" | sha256sum | cut -c1-16)"
+          measure() {
+            local new_image="$1" base_image="$2" variant="$3" out="$4"
+            local new_bytes current_bytes
+            new_bytes="$(docker image inspect "$new_image" --format '{{.Size}}')"
+            if docker pull --platform linux/amd64 "$base_image" >/dev/null 2>&1; then
+              current_bytes="$(docker image inspect "$base_image" --format '{{.Size}}')"
+            else
+              current_bytes=""
+              echo "No published baseline image found for ${base_image}"
+            fi
+            jq -n \
+              --argjson matrix "$MATRIX_JSON" \
+              --arg base "$base_image" \
+              --arg variant "$variant" \
+              --arg current "$current_bytes" \
+              --arg new "$new_bytes" \
+              '{matrix: $matrix, baseImage: $base, variant: $variant, currentBytes: $current, newBytes: $new}' \
+              > "size-report/${out}.json"
+          }
+          measure "$NEW_IMAGE" "$BASE_IMAGE" "" "$id"
+          measure "$NEW_SLIM_IMAGE" "$BASE_SLIM_IMAGE" "slim" "${id}-slim"
+          echo "id=${id}" >> "$GITHUB_OUTPUT"
+
+      - name: Upload image size report
+        if: github.event_name == 'pull_request'
+        uses: actions/upload-artifact@v7
+        with:
+          name: image-size-${{ steps.image-size.outputs.id }}
+          path: size-report/
+          retention-days: 1
+
       - name: Login to DockerHub
         if: github.event_name != 'pull_request'
         uses: docker/login-action@v4
@@ -319,3 +363,50 @@ jobs:
           cache-from: |
             type=gha,scope=${{ matrix.image-name }}-${{ matrix.node-version }}-${{ matrix.playwright-version }}-slim
             type=gha,scope=${{ matrix.image-name }}-${{ matrix.node-version }}-${{ matrix.playwright-version }}
+
+  # Aggregate the per-image size reports uploaded by the build matrix and post/update
+  # a single sticky PR comment comparing current vs new image sizes.
+  size-report:
+    name: Report image size changes
+    needs: [build-main]
+    if: ${{ always() && github.event_name == 'pull_request' }}
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    # All image workflows update sections of ONE shared PR comment; the repo-wide
+    # concurrency group serializes their read-modify-write so no section is lost.
+    concurrency:
+      group: image-size-comment-${{ github.event.pull_request.number }}
+      cancel-in-progress: false
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+
+      - name: Download image size reports
+        uses: actions/download-artifact@v8
+        continue-on-error: true
+        with:
+          pattern: image-size-*
+          path: size-reports
+          merge-multiple: true
+
+      - name: Set up Node.js
+        uses: actions/setup-node@v6
+        with:
+          node-version: 24
+
+      - name: Render comment section
+        run: |
+          mkdir -p size-reports
+          node .github/scripts/format-image-size-report.ts size-reports "Node + Playwright images" "${{ github.event.pull_request.head.sha }}" | tee image-size-comment.md
+
+      - name: Post or update image size comment
+        uses: actions/github-script@v8
+        env:
+          SECTION_KEY: node-playwright
+        with:
+          script: |
+            const { readFileSync } = require('node:fs');
+            const upsertSection = require('./.github/scripts/upsert-image-size-comment.js');
+            await upsertSection({ github, context }, process.env.SECTION_KEY, readFileSync('image-size-comment.md', 'utf8'));

.github/workflows/release-node-puppeteer.yaml

@@ -258,6 +258,50 @@ jobs:
       - name: Test slim image
         run: docker run ${{ fromJson(steps.prepare-slim-tags.outputs.result).firstImageName }}
 
+      - name: Measure image size
+        id: image-size
+        if: github.event_name == 'pull_request'
+        env:
+          MATRIX_JSON: ${{ toJSON(matrix) }}
+          NEW_IMAGE: ${{ fromJson(steps.prepare-tags.outputs.result).firstImageName }}
+          BASE_IMAGE: apify/actor-${{ matrix.image-name }}:${{ matrix.node-version }}
+          NEW_SLIM_IMAGE: ${{ fromJson(steps.prepare-slim-tags.outputs.result).firstImageName }}
+          BASE_SLIM_IMAGE: apify/actor-${{ matrix.image-name }}:${{ matrix.node-version }}-slim
+        run: |
+          set -euo pipefail
+          mkdir -p size-report
+          id="$(printf '%s' "$MATRIX_JSON" | sha256sum | cut -c1-16)"
+          measure() {
+            local new_image="$1" base_image="$2" variant="$3" out="$4"
+            local new_bytes current_bytes
+            new_bytes="$(docker image inspect "$new_image" --format '{{.Size}}')"
+            if docker pull --platform linux/amd64 "$base_image" >/dev/null 2>&1; then
+              current_bytes="$(docker image inspect "$base_image" --format '{{.Size}}')"
+            else
+              current_bytes=""
+              echo "No published baseline image found for ${base_image}"
+            fi
+            jq -n \
+              --argjson matrix "$MATRIX_JSON" \
+              --arg base "$base_image" \
+              --arg variant "$variant" \
+              --arg current "$current_bytes" \
+              --arg new "$new_bytes" \
+              '{matrix: $matrix, baseImage: $base, variant: $variant, currentBytes: $current, newBytes: $new}' \
+              > "size-report/${out}.json"
+          }
+          measure "$NEW_IMAGE" "$BASE_IMAGE" "" "$id"
+          measure "$NEW_SLIM_IMAGE" "$BASE_SLIM_IMAGE" "slim" "${id}-slim"
+          echo "id=${id}" >> "$GITHUB_OUTPUT"
+
+      - name: Upload image size report
+        if: github.event_name == 'pull_request'
+        uses: actions/upload-artifact@v7
+        with:
+          name: image-size-${{ steps.image-size.outputs.id }}
+          path: size-report/
+          retention-days: 1
+
       - name: Login to DockerHub
         if: github.event_name != 'pull_request'
         uses: docker/login-action@v4
@@ -299,3 +343,50 @@ jobs:
           cache-from: |
             type=gha,scope=${{ matrix.image-name }}-${{ matrix.node-version }}-${{ matrix.puppeteer-version }}-slim
             type=gha,scope=${{ matrix.image-name }}-${{ matrix.node-version }}-${{ matrix.puppeteer-version }}
+
+  # Aggregate the per-image size reports uploaded by the build matrix and post/update
+  # a single sticky PR comment comparing current vs new image sizes.
+  size-report:
+    name: Report image size changes
+    needs: [build-main]
+    if: ${{ always() && github.event_name == 'pull_request' }}
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    # All image workflows update sections of ONE shared PR comment; the repo-wide
+    # concurrency group serializes their read-modify-write so no section is lost.
+    concurrency:
+      group: image-size-comment-${{ github.event.pull_request.number }}
+      cancel-in-progress: false
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+
+      - name: Download image size reports
+        uses: actions/download-artifact@v8
+        continue-on-error: true
+        with:
+          pattern: image-size-*
+          path: size-reports
+          merge-multiple: true
+
+      - name: Set up Node.js
+        uses: actions/setup-node@v6
+        with:
+          node-version: 24
+
+      - name: Render comment section
+        run: |
+          mkdir -p size-reports
+          node .github/scripts/format-image-size-report.ts size-reports "Node + Puppeteer images" "${{ github.event.pull_request.head.sha }}" | tee image-size-comment.md
+
+      - name: Post or update image size comment
+        uses: actions/github-script@v8
+        env:
+          SECTION_KEY: node-puppeteer
+        with:
+          script: |
+            const { readFileSync } = require('node:fs');
+            const upsertSection = require('./.github/scripts/upsert-image-size-comment.js');
+            await upsertSection({ github, context }, process.env.SECTION_KEY, readFileSync('image-size-comment.md', 'utf8'));