Purge cache via API

POST /v1/delivery/purge/requests with the group id and media paths

One call purges: identify the domain, name the group, and — for path-based groups — list the URLs or tags. The response reports exactly what happened, including anything refused.

Purge logs table with date filters showing recent purge requests, actor, and status — Purge Logs records every purge with its scope, actor, and outcome.

The request#

curl -X POST -H "X-Api-Key: $TENANT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "domain": "www.acme.com",
    "purgeGroupId": "specific-urls",
    "mediaPaths": [
      "https://www.acme.com/products",
      "/assets/logo.png"
    ]
  }' \
  https://api.example.com/v1/delivery/purge/requests

domain (or domainId) resolves as in the options call; purgeGroupId (or groupId) is required (purgeGroupId is required.); paths travel in mediaPaths (or urls). Validation runs before anything is purged: at most 10 paths (At most 10 URLs can be purged per request.), no wildcards (A valid non-wildcard URL or path is required.), each path must belong to the domain (URL must belong to the selected domain.), and URL groups need at least one entry. full-site rejects paths outright — media paths are only supported for specific URL purge groups.

The response#

Success:

{
  "ok": true,
  "status": "COMPLETED",
  "requestId": "...",
  "group": { "id": "specific-urls", "label": "Specific URLs" },
  "purged": 14
}

If any path is refused the call answers 400 with "error": "Purge request rejected.", the same status (REJECTED) and requestId, plus rejectedPaths explaining each refusal — nothing is purged when even one path fails validation. Matching is exact: a URL is purged with its query string when one is present, and path-wide when it is not. Every API purge is recorded as activity cache.api_purge_requested (group, purged count, rejected count, request id) and lands in Cache → Purge logs with Request, URLs, Actor, Status, and Requested columns (cache-purge-logs screenshot below).

Expected result#

A request id you can correlate with the purge log, a purged count for observability, and honest rejection details rather than silent partial success.

Back to Purge cache via API