Use cases — databricks runtime 104 lts eol inventory
Two concrete scenarios for the 2026-11-01 migration window. Local-only helper; not affiliated with the vendor. 官方 EoS 表线性；缺「spark_version 含 10.4」多 workspace 残留盘点+owner 导出。不做 Databricks 控制台 SaaS。
Databricks platform admin: Before Nov 1, clear all workspaces of 10.4-pinned compute
Who / role: Databricks platform admin acting as databricks-platform-admin in context dbr-104-lts-eol-nov1. Expertise assumed: intermediate. Market frame: Databricks Runtime 10.4 LTS EOL Nov1 2026 — clusters/jobs still pinned to spark_version 10.4 (data platform Runtime domain).
Task: Before Nov 1, clear all workspaces of 10.4-pinned compute before deadline 2026-11-01. Vehicle is inventory-board; output expected is spark_version 10.4 residue inventory + Markdown/TSV export.
Steps: System table scan → IaC grep → owner assign → upgrade → re-verify. Use the home Hero to capture evidence locally, then paste the Markdown export into the migration ticket with official source URLs.
Result: No spark_version contains 10.4; ticket has exported inventory. Operators should keep a dated export attached to the change ticket, assign an accountable owner on every residue row, and re-open the official vendor documentation whenever a date or field name is unclear. Do not treat this local helper as a substitute for the vendor console or for production cutover approvals. Prefer verified primary sources over screenshots or chat summaries when communicating status to stakeholders.
Data engineering lead: Map critical jobs off 10.4 with Spark behavior retests
Who / role: Data engineering lead acting as databricks-platform-admin in context dbr-104-lts-eol-nov1. Expertise assumed: intermediate. Market frame: Databricks Runtime 10.4 LTS EOL Nov1 2026 — clusters/jobs still pinned to spark_version 10.4 (data platform Runtime domain).
Task: Map critical jobs off 10.4 with Spark behavior retests before deadline 2026-11-01. Vehicle is inventory-board; output expected is spark_version 10.4 residue inventory + Markdown/TSV export.
Steps: List jobs → pick target LTS → stage upgrade → compare outputs. Use the home Hero to capture evidence locally, then paste the Markdown export into the migration ticket with official source URLs.
Result: Critical jobs run on supported LTS with signed-off diffs. Operators should keep a dated export attached to the change ticket, assign an accountable owner on every residue row, and re-open the official vendor documentation whenever a date or field name is unclear. Do not treat this local helper as a substitute for the vendor console or for production cutover approvals. Prefer verified primary sources over screenshots or chat summaries when communicating status to stakeholders.
Next: home tool · how-to · FAQ.