4 - Automated CKAN Jobs #50113
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| --- | |
| name: 4 - Automated CKAN Jobs | |
| on: | |
| schedule: | |
| - cron: '30 7 * * *' # Tracking Update -- every day at 2:30am EST | |
| - cron: '0 2 * * *' # S3 Sitemap Update -- every day at 10pm EST | |
| - cron: '4/15 * * * *' # Harvester Check -- every 15 mins | |
| - cron: '0 3 * * *' # DB-Solr-Sync -- every day at 10pm EST | |
| - cron: '30 6 * * *' # Check Stuck Jobs -- every day at 1:30am EST | |
| - cron: '0 13 * * *' # Generate Reports -- every day @9:00am EST | |
| env: | |
| ERROR: false | |
| # Make sure 'schedule-cron' matches these varaibles. | |
| SCHEDULE_TRACKING: '30 7 * * *' | |
| SCHEDULE_SITEMAP: '0 2 * * *' | |
| SCHEDULE_HARVESTING: '4/15 * * * *' | |
| SCHEDULE_DBSOLR_SYNC: '0 3 * * *' | |
| SCHEDULE_STUCK_JOBS: '30 6 * * *' | |
| SCHEDULE_GEN_REPORT: '0 13 * * *' | |
| jobs: | |
| setup-matrix: | |
| runs-on: ubuntu-latest | |
| outputs: | |
| matrix: ${{ steps.generate-matrix.outputs.matrix }} | |
| steps: | |
| - name: Generate Matrix | |
| id: generate-matrix | |
| run: | | |
| # Matrix Configuration Settings | |
| # - Default App to run tasks on: cataog-admin | |
| # - Default timeout for task to raise issue: 6 hours | |
| # - Default Create an Information Issue for jobs we want to inspect: false | |
| # - Default monitor state: true | |
| # schedule: tracking update | |
| # - Only run on [development, staging, prod] | |
| # - Create Error issue: if runtime longer than 210 mins | |
| # - Create Informational issue: tracking-update-info.md | |
| # schedule: sitemap update | |
| # - Only run on [prod] | |
| # - Run on [catalog-gather] app | |
| # schedule: harvesting update | |
| # - Only run on [development, staging, prod] | |
| # schedule: stuck jobs check | |
| # - Only run on [staging, prod] | |
| # - Create Error issue: if >0 stuck jobs, automated_ckan_error.md | |
| # schedule: db-solr-sync | |
| # - Only run on [staging, prod] | |
| # - Create Error issue: if runtime longer than 30 mins | |
| # - Create Informational issue: db-solr-sync-info.md | |
| # schedule: report generation | |
| # - Only run on [development, staging, prod] | |
| # - Create Error issue: if ckan command errors | |
| MATRIX=$(cat << MAT | |
| { | |
| "schedule": ["${{env.SCHEDULE_TRACKING}}", "${{env.SCHEDULE_SITEMAP}}", | |
| "${{env.SCHEDULE_HARVESTING}}", "${{env.SCHEDULE_STUCK_JOBS}}", | |
| "${{env.SCHEDULE_DBSOLR_SYNC}}", "${{ env.SCHEDULE_GEN_REPORT}}"], | |
| "environ": ["development", "staging", "prod"], | |
| "include": [ {"app": "catalog-admin"}, | |
| {"error_seconds": 22000}, | |
| {"info_issue": false}, | |
| {"issue_template": ".github/automated_ckan_error.md"}, | |
| {"monitor": true}, | |
| {"schedule": "${{env.SCHEDULE_HARVESTING}}", "command": "ckan harvester run"}, | |
| {"schedule": "${{env.SCHEDULE_TRACKING}}", "command": "ckan geodatagov tracking-update"}, | |
| {"schedule": "${{env.SCHEDULE_TRACKING}}", "error_seconds": 12600}, | |
| {"schedule": "${{env.SCHEDULE_TRACKING}}", "info_issue": true}, | |
| {"schedule": "${{env.SCHEDULE_TRACKING}}", "issue_template": ".github/tracking-update-info.md"}, | |
| {"schedule": "${{env.SCHEDULE_SITEMAP}}", "command": "ckan geodatagov sitemap-to-s3"}, | |
| {"schedule": "${{env.SCHEDULE_SITEMAP}}", "app": "catalog-gather"}, | |
| {"schedule": "${{env.SCHEDULE_DBSOLR_SYNC}}", "command": "ckan geodatagov db-solr-sync"}, | |
| {"schedule": "${{env.SCHEDULE_DBSOLR_SYNC}}", "error_seconds": 1800}, | |
| {"schedule": "${{env.SCHEDULE_DBSOLR_SYNC}}", "info_issue": true}, | |
| {"schedule": "${{env.SCHEDULE_DBSOLR_SYNC}}", "issue_template": ".github/db-solr-sync-info.md"}, | |
| {"schedule": "${{env.SCHEDULE_STUCK_JOBS}}", "command": "ckan geodatagov check-stuck-jobs"}, | |
| {"schedule": "${{env.SCHEDULE_GEN_REPORT}}", "command": "ckan report generate"}, | |
| {"schedule": "${{env.SCHEDULE_GEN_REPORT}}", "issue_template": ".github/report-generation-failure.md"}, | |
| {"environ": "development", "ram": "1G"}, | |
| {"environ": "staging", "ram": "2500M"}, | |
| {"environ": "prod", "ram": "3G"} | |
| ], | |
| "exclude": [ {"schedule": "${{env.SCHEDULE_SITEMAP}}", "environ": "development"}, | |
| {"schedule": "${{env.SCHEDULE_SITEMAP}}", "environ": "staging"}, | |
| {"schedule": "${{env.SCHEDULE_DBSOLR_SYNC}}", "environ": "development"}, | |
| {"schedule": "${{env.SCHEDULE_STUCK_JOBS}}", "environ": "development"} | |
| ], | |
| } | |
| MAT | |
| ) | |
| echo "matrix<<EOF" >> $GITHUB_OUTPUT | |
| echo "$MATRIX" >> $GITHUB_OUTPUT | |
| echo "EOF" >> $GITHUB_OUTPUT | |
| ckan-auto-command: | |
| needs: | |
| - setup-matrix | |
| strategy: | |
| max-parallel: 1 | |
| fail-fast: false | |
| matrix: ${{ fromJSON(needs.setup-matrix.outputs.matrix) }} | |
| name: ${{ matrix.command }} - ${{matrix.environ}} | |
| runs-on: ubuntu-latest | |
| environment: ${{matrix.environ}} | |
| steps: | |
| - name: checkout | |
| uses: actions/checkout@v4 | |
| - name: Set Unique Command Name | |
| run: | | |
| INSTANCE_NAME="$(echo ${{ matrix.command }}-${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT} | tr -d ' ')" | |
| echo "INSTANCE_NAME=${INSTANCE_NAME}" | tee -a $GITHUB_ENV | |
| echo "START=`date +%s`" | tee -a $GITHUB_ENV | |
| - name: ${{ matrix.command }} | |
| if: ${{ github.event.schedule == matrix.schedule }} | |
| uses: cloud-gov/cg-cli-tools@main | |
| with: | |
| command: > | |
| cf run-task ${{ matrix.app }} | |
| --command "${{ matrix.command }}" | |
| --name "$INSTANCE_NAME" | |
| -m ${{ matrix.ram }} -k 2G | |
| cf_org: gsa-datagov | |
| cf_space: ${{ matrix.environ }} | |
| cf_username: ${{secrets.CF_SERVICE_USER}} | |
| cf_password: ${{secrets.CF_SERVICE_AUTH}} | |
| - name: monitor task output | |
| if: ${{ matrix.monitor && github.event.schedule == matrix.schedule }} | |
| uses: cloud-gov/cg-cli-tools@main | |
| with: | |
| command: > | |
| tools/monitor_cf_logs.sh ${{ matrix.app }} $INSTANCE_NAME | |
| cf_org: gsa-datagov | |
| cf_space: ${{ matrix.environ }} | |
| cf_username: ${{secrets.CF_SERVICE_USER}} | |
| cf_password: ${{secrets.CF_SERVICE_AUTH}} | |
| - name: check runtime | |
| run: | | |
| END=`date +%s` | |
| ELAPSED=$(( END - START)) | |
| echo "$START, $END, $ELAPSED" | |
| if [[ $ELAPSED > ${{ matrix.error_seconds }} ]]; then echo ERROR=true | tee -a $GITHUB_ENV; fi; | |
| if [[ $ELAPSED > ${{ matrix.error_seconds }} ]]; then echo "ELAPSED=${ELAPSED}" | tee -a $GITHUB_ENV; fi; | |
| - name: Create Issue for auditing 📑 or failure 😢 | |
| if: ${{ failure() || (env.ERROR == true) || (matrix.info_issue && github.event.schedule == matrix.schedule) }} | |
| uses: JasonEtco/create-an-issue@v2 | |
| env: | |
| GITHUB_TOKEN: ${{ secrets.ADD_TO_PROJECT_PAT }} | |
| GITHUB_JOB: ${{ toJson(github)['job'] }} | |
| GITHUB_ATTEMPTS: ${{ github.run_attempt }} | |
| LAST_COMMIT: ${{ github.sha }} | |
| LAST_RUN_BY: ${{ github.actor }} | |
| RUN_ID: ${{ github.run_id }} | |
| COMMAND: ${{ matrix.command }} | |
| SCHEDULE: ${{ matrix.schedule }} | |
| ENVIRONMENT: ${{ matrix.environ }} | |
| EXEC_TIME: ${{ env.ELAPSED }} | |
| with: | |
| filename: ${{ matrix.issue_template }} | |
| update_existing: true |