Pre-warm Census syncs before the sales day starts

Every weekday at 6:45am, pre-warm the sales-critical Census syncs so revenue teams start the day with fresh warehouse data in the CRM and marketing tools, and drop a confirmation card in #sales-ops Slack.

Trigger: cron, Monday through Friday at 6:45am in the team's local timezone. This is a deterministic workflow (fixed sync IDs, fixed poll loop, fixed message template), so build it as a code workflow, not an agent.

Inputs to expose at the top of the workflow so they are easy to edit later:

1. A configured list of Census sync IDs to pre-warm. Seed the list with three placeholders labeled Salesforce accounts sync, HubSpot contact scores sync, and marketing suppression list sync so the user can paste in their actual sync IDs. 2. The Slack channel to notify (default #sales-ops). 3. The Google Sheet ID and tab name for the audit trail (default tab: Morning Sync Log). 4. A per-sync poll timeout (default 20 minutes) and poll interval (default 30 seconds).

For each sync ID in the list, in order:

1. Call census.Trigger Sync Run to start the run. Capture the returned sync_run_id. Important nuance from Census: this endpoint returns 200 when the run is queued, NOT when it succeeds. Do not infer success from the trigger response.

2. Poll census.Fetch Sync Run with that sync_run_id every 30 seconds until status is no longer "running". Typical terminal values are "completed" and "failed". Stop polling on any non-running status, or when the per-sync timeout is reached (record that sync as "timed_out" and move on so one hanging sync does not block the rest).

3. Collect a result row: sync_id, sync name (from the sync run response if available), final status, records_processed, records_failed, error_code and error_message (when status is "failed"), started_at and finished_at, and a direct Census URL to the run (app.getcensus.com/syncs/{sync_id}/runs/{sync_run_id}, or app-eu.getcensus.com for EU workspaces).

Once every sync in the list has reached a terminal state (or timed out), do the two writes:

A. Send one Slack message via slack.Send a Message to the configured channel. Format it as a single summary card, not one message per sync. Include: the date, total syncs run, how many completed vs failed vs timed_out, and one line per sync showing status emoji + sync name + records_processed (and records_failed if any). For every sync whose status is failed or timed_out, include the direct Census link so the on-call person can jump straight into the run. Use a green check for completed, red X for failed, and clock for timed_out.

B. For each sync result, append one row to the Morning Sync Log tab of the configured Google Sheet using google-sheets.Append Values. Columns in this order: run_date, sync_id, sync_name, status, records_processed, records_failed, duration_seconds, error_code, error_message, census_url. This gives a rolling audit trail.

Robustness rules: if census.Trigger Sync Run itself errors (400/401/403/404/422/429), record that sync as status "trigger_failed" with the error message, include it in the Slack summary and the sheet row, and keep going with the next sync — one broken sync should never abort the batch. Respect Retry-After on 429s from Census. Do not re-trigger a sync that is already running; if the trigger response indicates a run is already in progress, poll the existing run instead.

Community context: teams commonly schedule reverse-ETL syncs to fire just before the sales day starts so revenue tools always have fresh warehouse data by the time reps log in. This workflow codifies that pattern into a repeatable morning pre-warm with a visible receipt in Slack and an audit trail in Sheets.