Get one call's full internal lifecycle event log

One bounded page of this call's call_events rows, oldest first, each with its raw event payload. This is the platform's own internal bookkeeping log, not a webhook replay: it includes types never delivered to any webhook endpoint at all (e.g. wallet.reserved, call.authorization, call.policy_evaluated, number.matched), and every row here is this row's own id/type/payload, never the signed webhook envelope. call.failed is written here too (payload shape varies by writer -- see CallEvent.payload), always thinner than the richer result body its webhook delivers separately. call.completed is the one terminal type NEVER written here at all -- it is derived from the call's final status and delivered only to webhooks; look at status on GET /v1/calls/{call_id} for that outcome instead. The response body is the plain array it has always been; the paging state is in headers. Do NOT read one page as the whole history: check X-Has-More, and continue with after=<the last row's page_cursor> until it is false. A long-running call with many agent tool invocations (each is its own call.agent_tool_call row) is exactly the case that needs more than one page.