fix: increase packer snapshot transfer timeout to 60m (#2648)

* fix: increase packer snapshot transfer timeout to 60m

The default 30m timeout is too short for transferring snapshots to
distant DO regions (blr1, sgp1, syd1). This caused zeroclaw and
kilocode builds to fail despite successful provisioning.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* revert: remove batch splitting from packer workflow

DO droplet cap is no longer an issue — revert to single parallel build
job for all agents.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

.github/workflows/packer-snapshots.yml

@@ -19,160 +19,29 @@ jobs:
     name: Generate matrix
     runs-on: ubuntu-latest
     outputs:
-      batch1: ${{ steps.set.outputs.batch1 }}
-      batch2: ${{ steps.set.outputs.batch2 }}
+      agents: ${{ steps.set.outputs.agents }}
     steps:
       - uses: actions/checkout@v4
       - id: set
         run: |
           SINGLE_AGENT="${SINGLE_AGENT_INPUT}"
           if [ -n "$SINGLE_AGENT" ]; then
-            # Single agent mode — put it in batch1 only
-            echo "batch1=[\"${SINGLE_AGENT}\"]" >> "$GITHUB_OUTPUT"
-            echo "batch2=[]" >> "$GITHUB_OUTPUT"
+            echo "agents=[\"${SINGLE_AGENT}\"]" >> "$GITHUB_OUTPUT"
           else
-            # Split agents into 2 batches to stay under DO's concurrent droplet cap
             AGENTS=$(jq -c 'keys' packer/agents.json)
-            TOTAL=$(echo "$AGENTS" | jq 'length')
-            HALF=$(( (TOTAL + 1) / 2 ))
-            BATCH1=$(echo "$AGENTS" | jq -c ".[:${HALF}]")
-            BATCH2=$(echo "$AGENTS" | jq -c ".[${HALF}:]")
-            echo "batch1=${BATCH1}" >> "$GITHUB_OUTPUT"
-            echo "batch2=${BATCH2}" >> "$GITHUB_OUTPUT"
+            echo "agents=${AGENTS}" >> "$GITHUB_OUTPUT"
           fi
         env:
           SINGLE_AGENT_INPUT: ${{ inputs.agent }}
 
-  batch1:
+  build:
     name: "Build ${{ matrix.agent }}"
     needs: matrix
-    if: ${{ needs.matrix.outputs.batch1 != '[]' }}
     runs-on: ubuntu-latest
     strategy:
       fail-fast: false
       matrix:
-        agent: ${{ fromJson(needs.matrix.outputs.batch1) }}
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: Read agent config
-        id: config
-        run: |
-          TIER=$(jq -r --arg a "$AGENT_NAME" '.[$a].tier // "minimal"' packer/agents.json)
-          INSTALL=$(jq -c --arg a "$AGENT_NAME" '.[$a].install // []' packer/agents.json)
-          echo "tier=${TIER}" >> "$GITHUB_OUTPUT"
-          echo "install=${INSTALL}" >> "$GITHUB_OUTPUT"
-        env:
-          AGENT_NAME: ${{ matrix.agent }}
-
-      - name: Setup Packer
-        uses: hashicorp/setup-packer@main
-        with:
-          version: latest
-
-      - name: Init Packer plugins
-        run: packer init packer/digitalocean.pkr.hcl
-
-      - name: Generate variables file
-        run: |
-          jq -n \
-            --arg token "$DO_API_TOKEN" \
-            --arg agent "$AGENT_NAME" \
-            --arg tier "$TIER" \
-            --argjson install "$INSTALL_COMMANDS" \
-            '{
-              do_api_token: $token,
-              agent_name: $agent,
-              cloud_init_tier: $tier,
-              install_commands: $install
-            }' > packer/auto.pkrvars.json
-        env:
-          DO_API_TOKEN: ${{ secrets.DO_API_TOKEN }}
-          AGENT_NAME: ${{ matrix.agent }}
-          TIER: ${{ steps.config.outputs.tier }}
-          INSTALL_COMMANDS: ${{ steps.config.outputs.install }}
-
-      - name: Build snapshot
-        run: packer build -var-file=packer/auto.pkrvars.json packer/digitalocean.pkr.hcl
-
-      - name: Cleanup old snapshots
-        if: success()
-        run: |
-          # DO snapshots don't support tags — filter by name prefix instead
-          PREFIX="spawn-${AGENT_NAME}-"
-          SNAPSHOTS=$(curl -s -H "Authorization: Bearer ${DO_API_TOKEN}" \
-            "https://api.digitalocean.com/v2/images?private=true&per_page=100" \
-            | jq -r --arg prefix "$PREFIX" \
-              '[.images[] | select(.name | startswith($prefix))] | sort_by(.created_at) | reverse | .[1:] | .[].id')
-
-          for ID in $SNAPSHOTS; do
-            echo "Deleting old snapshot: ${ID}"
-            curl -s -X DELETE -H "Authorization: Bearer ${DO_API_TOKEN}" \
-              "https://api.digitalocean.com/v2/images/${ID}" || true
-          done
-        env:
-          DO_API_TOKEN: ${{ secrets.DO_API_TOKEN }}
-          AGENT_NAME: ${{ matrix.agent }}
-
-      - name: Submit to DO Marketplace
-        if: success()
-        run: |
-          # Skip if no marketplace app IDs configured
-          if [ -z "$MARKETPLACE_APP_IDS" ]; then
-            echo "No MARKETPLACE_APP_IDS secret — skipping marketplace submission"
-            exit 0
-          fi
-
-          # Look up this agent's app ID from the JSON map
-          APP_ID=$(echo "$MARKETPLACE_APP_IDS" | jq -r --arg a "$AGENT_NAME" '.[$a] // empty')
-          if [ -z "$APP_ID" ]; then
-            echo "No marketplace app ID for agent ${AGENT_NAME} — skipping"
-            exit 0
-          fi
-
-          # Extract snapshot ID from Packer manifest
-          # artifact_id format is "region:snapshot_id" (e.g. "sfo3:12345678")
-          IMG_ID=$(jq '.builds[-1].artifact_id | split(":")[1] | tonumber' packer/manifest.json)
-          if [ -z "$IMG_ID" ] || [ "$IMG_ID" = "null" ]; then
-            echo "Failed to extract snapshot ID from manifest"
-            exit 1
-          fi
-
-          echo "Submitting snapshot ${IMG_ID} for ${AGENT_NAME} (app: ${APP_ID})"
-
-          # PATCH the Vendor API — updates go to "pending" review.
-          # 400 = app already pending/in-review (expected for nightly runs), not an error.
-          HTTP_CODE=$(curl -s -o /tmp/mp-response.json -w "%{http_code}" \
-            -X PATCH \
-            -H "Content-Type: application/json" \
-            -H "Authorization: Bearer ${DO_API_TOKEN}" \
-            -d "$(jq -n \
-              --arg reason "Nightly rebuild — $(date -u '+%Y-%m-%d')" \
-              --argjson imageId "$IMG_ID" \
-              '{reasonForUpdate: $reason, imageId: $imageId}')" \
-            "https://api.digitalocean.com/api/v1/vendor-portal/apps/${APP_ID}")
-
-          case "$HTTP_CODE" in
-            200) echo "Marketplace submission accepted (pending review)" ;;
-            400) echo "App already pending review — skipping (expected for nightly runs)" ;;
-            *)   echo "Marketplace API returned ${HTTP_CODE}:"
-                 cat /tmp/mp-response.json
-                 exit 1 ;;
-          esac
-        env:
-          DO_API_TOKEN: ${{ secrets.DO_API_TOKEN }}
-          AGENT_NAME: ${{ matrix.agent }}
-          MARKETPLACE_APP_IDS: ${{ secrets.MARKETPLACE_APP_IDS }}
-
-  batch2:
-    name: "Build ${{ matrix.agent }}"
-    needs: [matrix, batch1]
-    if: ${{ needs.matrix.outputs.batch2 != '[]' && always() }}
-    runs-on: ubuntu-latest
-    strategy:
-      fail-fast: false
-      matrix:
-        agent: ${{ fromJson(needs.matrix.outputs.batch2) }}
+        agent: ${{ fromJson(needs.matrix.outputs.agents) }}
     steps:
       - uses: actions/checkout@v4
 

packer/digitalocean.pkr.hcl

@@ -45,6 +45,9 @@ source "digitalocean" "spawn" {
     "nyc1", "nyc3", "sfo3", "tor1", "ams3",
     "lon1", "fra1", "blr1", "sgp1", "syd1",
   ]
+
+  # Default is 30m which times out for distant regions (blr1, sgp1, syd1)
+  transfer_timeout = "60m"
 }
 
 build {