Pagination

Many vManage endpoints (alarms, events, bulk statistics, large inventory listings) return one page of a larger result set. The server detects two pagination shapes from the OpenAPI spec and auto-follows them, returning a single stitched response.

Shapes

• scroll — Elasticsearch-style. Request param scrollId; response carries pageInfo.scrollId and pageInfo.hasMoreData. Used by alarms, events, bulk stats.
• offset — Classic page / pageSize (or count / limit). Used by most configuration and inventory listings.

Behaviour

When pagination is engaged the response is wrapped:

{
  "data": [ ...concatenated items from every page... ],
  "pagination": {
    "style": "scroll",
    "pages_fetched": 5,
    "truncated": true,
    "next_cursor": {"scrollId": "DXF1Z..."}
  }
}

• truncated: true ⇔ next_cursor is non-null. To get the next batch, call the same action again with next_cursor's fields merged into params — e.g. params["scrollId"] = "DXF1Z..." (scroll) or params["page"] = 6 (offset).
• Non-paginated actions and explicit opt-outs return the raw response shape unchanged.

Configuration

sdwan:
  pagination:
    enabled: true     # master switch
    max_pages: 5      # default cap on auto-follow
    page_size: null   # null → use the endpoint's natural page size

Per-call overrides

The dispatcher recognises three reserved underscore keys in params. They are stripped before the request is sent to vManage:

Param	Type	Effect
_max_pages	int	Override max_pages for this call.
_page_size	int	Override page_size for this call (offset only).
_pagination	"off"	Skip pagination; return the raw first page. The literal "off" is the only recognised disabling value.

Examples

Fetch up to 10 pages of alarms:

{ "action": "get_alarms", "params": { "_max_pages": 10 } }

Disable auto-follow for one call:

{ "action": "get_alarms", "params": { "_pagination": "off" } }

Resume from a returned cursor:

{ "action": "get_alarms", "params": { "scrollId": "DXF1Z..." } }