Fivetran package v1.0.0

List the account's destinations; this pack registers the collection read only — single-destination reads are not registered. Destination records reveal where the organization lands its replicated data (service, region, group ids), which is internal infrastructure information.

The host is a reserved placeholder: substitute api.fivetran.com before enabling.

List the available connector types and their internal service identifiers and configuration schemas — vendor catalog information used to build connection requests, classified as internal data because the response drives downstream provisioning decisions. The registered path is the current '/v1/metadata/connector-types' shown on the fetched metadata-connectors reference page (the legacy '/v1/metadata/connectors' path was sunset in 2022).

The host is a reserved placeholder: substitute api.fivetran.com before enabling.

List the account's connections (List All Connections); this pack registers the collection read only — Retrieve Connection Details (GET /v1/connections/{connectionId}) is deliberately not registered because the fetched connection-details page's response example echoes a credential-valued config field — the example does not prove live responses return unmasked secrets, but the detail response must be treated as credential-bearing until proven otherwise, and a read that may return live secrets needs a response-handling review this pack has not done.

Connection records reveal which systems replicate into which destination, their schedules and their sync state — internal pipeline information.

The host is a reserved placeholder: substitute api.fivetran.com before enabling.

Update one connection's mutable fields — pausing and resuming (paused), the scheduling method (schedule_type auto or manual; setting manual makes every subsequent sync occur only through an explicit trigger call), the sync frequency and the networking settings. Resuming a paused connection or setting schedule_type to auto re-enables provider-side execution: Fivetran's scheduler then moves customer data on its own without traversing fivetran.connection.sync.trigger, so the update carries the deploy-class effect and the capability declares update and deploy at the minimum R3 tier.

This registration does not restrict which fields the request body may carry: the body can also carry config values, which for many services are source credentials, so a request body can itself contain live secrets and must be handled like one. A reviewer who intends pause-only must impose that in policy, because intended limits need a separate enforceable control. CONNECTION_ID is a placeholder the deployment binds before enabling.

Trigger a data sync for one connection without waiting for its schedule. Registered with the deploy action at the minimum R3 tier rather than as an ordinary execution: the sync moves customer data from the source system into the destination, overwriting destination state, and that data movement is not reversed by a later call. Deploy is one of the designated actions requiring human approval.

On supported execution paths mediated by AAES Govern, irreversible actions require approval by an authorized person; this does not prevent calls through bypass or observation-only paths.

Request body: the documented body is {"force": true|false}; the fetched sync-connection page confirms force=true 'will stop the sync and re-run it' — terminating an in-flight sync is a terminate-class, irreversible interruption (the interrupted sync's partial work is not rolled back), so the delete action is declared alongside deploy; an empty body syncs only if the connection is idle. The registration does not restrict the body, so the interrupt-in-flight behavior is in scope. CONNECTION_ID is a placeholder the deployment binds before enabling.

Trigger a full historical re-sync of a connection. Registered with the deploy action at the minimum R3 tier: a re-sync re-reads the source from the beginning and rewrites the destination tables — heavier than a sync on both the source system (full re-extract) and the destination (rewritten tables), and it is not reversed by a later call. Deploy is one of the designated actions requiring human approval.

Request body: the documented body is an optional scope object mapping schema names to non-empty arrays of table names; omitting it re-syncs every schema and table of the connection, which is the default a caller gets with an empty body. The API declines with 409 while another sync is in progress and re-syncs a paused connection when it is re-enabled — documented timing behaviors, not restrictions this pack imposes. CONNECTION_ID is a placeholder the deployment binds before enabling.

Delete a connection from the Fivetran account. The pipeline definition — source configuration, sync cursor, schedule — is removed; recreating the connection later starts as a new pipeline (a new initial sync), not an undo of this call, and any downstream models or dashboards fed by the connection stop updating. Registered as delete at the minimum R3 tier; delete is one of the designated actions requiring human approval. CONNECTION_ID is a placeholder the deployment binds before enabling.