Statistics API query concepts
Authenticate against your tenant API base URL using an API token with permission for the requested operation. For the shared token header and environment setup, use Overview; keep credentials private. Send a JSON request body for POST queries. The method is part of each contract: POST can retrieve analytics without changing a resource.
Replace path parameters such as {namespace}, {name}, {site}, and {network_id} with the identifiers of
the selected resource. The paths on resource pages are complete API paths relative to the tenant base URL. Query
reference shows how to discover load balancer and virtual host
identities. A configuration object name, graph node identity, API endpoint path, and site name serve different
purposes; use the identifier required by the linked request schema.
| Query concept | How to use it |
|---|---|
| Namespace and tenant access | A namespace query normally selects that namespace. Use tenant-wide queries only when the operation documents them and your account has access. A literal system path, or a row marked system only, requires that reserved namespace. Some DDoS and site APIs use {namespace} in the path but accept only system. |
| Time window | Set the endpoint’s start and end fields explicitly. Many metrics and log APIs accept Unix timestamps or RFC 3339 timestamps; other contracts use timeRange, timestamp, period_in_days, or nested time filters. Consult the schema for defaults and supported formats. |
| Resolution and rollups | step or step_size controls point spacing. range controls the lookback used to evaluate each point. They serve different purposes. Stored rollups can increase the effective step; use the returned step and timestamps. Load balancer metrics shows the measured effect. |
| Metric selection | field_selector, metric_selector, or metrics selects returned measurements. Graph APIs separate node and edge metrics and health scores. Select a meaningful family and supported dimensions; the schema defines allowed combinations. |
| Filtering | Contracts use different selectors: label_filter, a matcher string filter, structured filters, or log query expressions. These are endpoint-specific formats. A virtual host filter accepted by one API does not automatically apply to another. |
| Grouping and aggregation | group_by determines dimensions; aggs requests named log/event aggregations. A top-N query returns a ranking, not every matching record. Choose supported grouping labels and aggregation types from the request schema. |
| Pagination | Follow the endpoint’s cursor, scroll, search-after, offset, or page contract. Several log and event APIs cap a batch at 500 records. An empty returned scroll_id ends scrolling. Compare collected records with total_hits where supplied; a first page is not a complete window. |
| Responses and empty data | Inspect the successful response schema, units, labels, timestamps, and returned resolution. HTTP 200 with no series means the selected window has no returned points. It does not establish zero traffic or populated telemetry. |
Reading the endpoint tables
Section titled “Reading the endpoint tables”Each method/path appears once under its primary resource group. Statistics and analytics queries retrieve measurements, logs, events, or analytical summaries. Related discovery and status queries help identify resources, supported data, and current state. Existing reports and artifacts retrieve previously available output. Configuration creation/replacement/deletion, enable/disable operations, credential retrieval, traffic generation, and report-generation requests are outside this reference.
The operation link explains the query, and the schema links define its request and successful response. Routes marked corrected source are documented in the corrected specifications but absent from the merged enriched specification. The three cloud-connect metric queries are retained with direct source links. A newer version does not imply deprecation of an older route; a deprecation label appears only when the source declares it.
For API endpoint paths ending in /pdf, the response is a probability density function of sizes or latency, not a PDF document download. Report download routes are listed separately.