Tuning Incremental Sync Config

A sync’s configuration lives on the ExternalDataSchema and can be changed any time via external-data-schemas-partial-update. Most changes are non-destructive (take effect on the next sync), but a few (switching sync_type, changing primary keys) require careful handling to avoid corrupting the synced data.

When to use this skill

• The user wants to change how an already-connected table is synced
• A diagnosis flagged the incremental field or primary key as wrong
• The table is syncing too often / not often enough
• Switching an incremental table to CDC (or vice versa)
• The source table was changed on the other side (new columns, dropped columns) and the sync config needs to catch up

If the user is setting up a brand-new source, use setting-up-a-data-warehouse-source instead — configuration is chosen at creation time there.

Available tools

The fields you can tune

From the partial-update endpoint:

Workflow

Step 1 — Read the current config

Always start with external-data-schemas-retrieve({id}). Understanding the current state prevents mistakes like “fixing” an incremental_field that’s actually correct.

Note:

• Current sync_type, incremental_field, incremental_field_type, primary_key_columns
• Current status (don’t tune a schema that’s currently Running — wait or cancel first)
• last_synced_at (so you can tell if the next sync worked)
• latest_error if present (the error often tells you exactly what to change)

Step 2 — If changing sync_type or incremental_field, refresh candidates

Call external-data-schemas-incremental-fields-create({id}). Even though the operation name says “create”, it re-reads the source and returns the current candidate fields — use it to confirm the field you want to set actually exists on the source and which sync types are now available for this table.

The response:

{
  "incremental_fields": [{"field": "updated_at", "type": "datetime", ...}, ...],
  "incremental_available": true,
  "append_available": true,
  "cdc_available": true,
  "full_refresh_available": true,
  "detected_primary_keys": ["id"],
  "available_columns": [...]
}

If your target incremental_field isn’t in the list, tell the user — they need to either pick a different field or change the source table to add one.

Step 3 — Apply the change

Call external-data-schemas-partial-update({id}, {...changed fields}).

Only send the fields that are actually changing. Partial update means unspecified fields stay as they are.

Examples:

// Switch from full_refresh to incremental
{
  "sync_type": "incremental",
  "incremental_field": "updated_at",
  "incremental_field_type": "datetime"
}

// Change sync frequency to hourly
{"sync_frequency": "1hour"}

// Fix wrong PK on a CDC table
{"primary_key_columns": ["tenant_id", "order_id"]}

// Pause a schema

Step 4 — Decide whether existing data is still valid

This is the step that’s easy to get wrong. Some config changes invalidate the synced data; others don’t.

Changes that DON’T invalidate existing data:

• sync_frequency, sync_time_of_day — scheduling only
• should_sync — on/off
• cdc_table_mode in most cases — next sync will start writing to the new shape, but historical consolidated rows stay valid
• Switching between incremental and full_refresh with the same incremental_field — next sync just re-runs fresh
• Switching to or from sync_type: "webhook" — the synced data stays valid; only the ingestion path changes. Remember to register or unregister the webhook (see sections below) alongside the sync_type change.

Changes that MAY invalidate existing data and need a resync:

• Changing incremental_field to a different column — the high-water mark is from the old column and won’t match. Without a resync you’ll miss rows that were updated between the two fields’ histories.
• Changing primary_key_columns — existing rows may be deduplicated incorrectly against new PK definitions.
• Switching from full_refresh to append — the existing rows don’t have the version-history shape that append expects.
• Switching from append to full_refresh — opposite problem; you’ll end up with duplicate historical versions.
• Switching to/from cdc — the table shape changes fundamentally.

When the change invalidates data, the clean flow is:

1. external-data-schemas-partial-update with the new config
2. Warn the user this is destructive
3. external-data-schemas-resync to wipe and re-import under the new config

Or equivalently, external-data-schemas-delete-data → external-data-schemas-reload. delete-data + reload is cleaner when the table is large and the user wants to start from zero.

Step 5 — Trigger and confirm

For non-destructive changes, call external-data-schemas-reload({id}) to pick up the new config immediately rather than waiting for the schedule.

Wait a moment, then external-data-schemas-retrieve({id}) to confirm status = Running then Completed. Report last_synced_at and any new latest_error.

Specific common changes

Switching full_refresh → incremental

1. incremental-fields-create to confirm the desired field exists and incremental_available: true.
2. partial-update: {sync_type: "incremental", incremental_field, incremental_field_type}.
3. No data wipe needed — next sync just switches strategy. If the source is growing fast, the next incremental sync is the cheap one.

Switching incremental → cdc (Postgres only)

1. Run external-data-sources-check-cdc-prerequisites-create on the parent source. Only proceed if valid: true.
2. incremental-fields-create to confirm cdc_available: true and see detected_primary_keys.
3. partial-update: {sync_type: "cdc", primary_key_columns: [...], cdc_table_mode: "consolidated"}.
4. Resync required — CDC tables have a different shape. Trigger external-data-schemas-resync after the update. Warn the user this wipes existing data.

Fixing a stale incremental field after schema drift

Source dropped the updated_at column. Sync has been failing with “column does not exist”.

1. incremental-fields-create to see what fields remain.
2. Pick a replacement (or switch to full_refresh if none are suitable).
3. partial-update with the new field + type (or new sync_type).
4. reload to retry.

Changing primary keys on a CDC table

1. partial-update: {primary_key_columns: [...]}.
2. Resync required — existing CDC tombstones and upsert keys won’t match the new PK definition, leading to row duplication or missed updates.
3. resync, warn the user.

Changing sync_frequency

1. partial-update: {sync_frequency: "1hour"}.
2. No reload needed — the next scheduled sync picks up the new cadence. Or reload manually if the user wants to confirm nothing broke.

Switching a schema to sync_type: "webhook"

Only works for sources that implement WebhookSource (today: Stripe) and tables where supports_webhooks: true from incremental-fields-create.

1. incremental-fields-create to confirm supports_webhooks: true for the table.
2. partial-update: {sync_type: "webhook"}.
3. If the source doesn’t already have a webhook registered (check with webhook-info-retrieve), call external-data-sources-create-webhook-create({source_id}) to register it.
4. No resync required — the schema’s existing bulk-synced data stays, and the webhook becomes the primary ingestion path once the next reconciliation finishes.
5. Keep sync_frequency set (e.g. 24hour) — it acts as a safety-net reconciliation in case any webhook delivery is missed.

Switching off sync_type: "webhook"

1. partial-update: {sync_type: "incremental"} (or whatever bulk type is appropriate) with the required incremental_field + incremental_field_type.
2. If no other schemas on the source are still using sync_type: "webhook", call external-data-sources-delete-webhook-create({source_id}) to unregister. Leaving an orphaned webhook registered on the source side just means events will be received and dropped — not harmful, but messy.
3. If other schemas on the source are still on webhook, leave the webhook registered — it’s shared across all webhook-type schemas on the source.

Rotating a webhook signing secret

The source’s signing secret (e.g. Stripe’s whsec_...) was rotated, and payloads are now failing signature verification.

1. Grab the new secret from the source’s dashboard.
2. external-data-sources-update-webhook-inputs-create({source_id}, {inputs: {signing_secret: "whsec_..."}}).
3. No reload needed — the next inbound webhook payload will verify against the new secret.

Pausing a schema

1. partial-update: {should_sync: false}. Schema stops syncing but stays configured.
2. To resume later: partial-update: {should_sync: true}, then reload for an immediate run.

Important notes

• Read before you write. Always retrieve the current config first. partial-update doesn’t complain if you set a field to the value it already had, but you might be about to change something you didn’t realize was already set.
• Not every sync_type is available on every schema. The incremental-fields-create response tells you what’s available right now, which can be different from what was available at creation (e.g. CDC may have been enabled for the team since).
• Wipe when the shape changes. Switching sync strategy often changes the physical table. If you don’t resync, you’ll be mixing row shapes and queries will return garbage.
• CDC needs prerequisites. Never switch to sync_type: "cdc" without running check-cdc-prerequisites-create first. The sync will just fail immediately.
• Don’t touch a Running schema. If the schema is currently running, either wait for it to finish or external-data-schemas-cancel before applying the change. Updating config mid-sync can leave the incremental high-water mark inconsistent.
• Sync frequency is cheap to change. Encourage experimentation there. Sync_type and incremental_field are expensive to change — encourage care.
• Webhooks are registered at the source level, not the schema level. Multiple webhook-type schemas on the same source share one webhook registration. Only delete the webhook when the last webhook-type schema on that source is being switched away, otherwise other schemas stop receiving pushes.