Late Table Load Watchdog Correlated to Pipeline Logs in Axiom

Instead of just announcing that a table is stale, this workflow joins the freshness miss to the actual pipeline execution logs. It detects a late Snowflake load, queries Axiom for the corresponding ingest job's log lines around the expected run time, and delivers a Slack alert that already contains the likely root cause (timeout, source API 500s, retries exhausted, etc.).

Use it when your ingestion runs are observable in Axiom and you want the on-call engineer to open Slack and immediately see *why* the load is late, not just *that* it is late.

1. 1A schedule triggers the freshness sweep.
2. 2It reads the latest load timestamp per monitored table from Snowflake.
3. 3A logic step flags any table whose load is later than its SLA.
4. 4For each late table, it runs an Axiom query scoped to that job's dataset and time window to extract error and warning lines.
5. 5It posts a Slack message to the data on-call channel with the table name, minutes late, and the relevant log excerpt.