Delivery
Cache
Understand how responses are cached, inspect variations, and purge content when it changes.
How caching works
GET and HEAD responses under 1 MB are cached, with TTL clamped from the origin's max-age header.
- Public, cacheable GET/HEAD responses only
- Payloads under 1 MB
- TTL clamped from origin headers
- Hit ratio tracked on the overview
Cache variations
Serve different cached responses for the same URL based on request conditions, with rules and TTLs scoped per domain.
- Vary by device, query, header, cookie, country, IP, or user agent
- Fixed buckets and else values for query rules
- Cache rules scoped per domain
- Cache records mirrored in the dashboard
Purge content
Clear cached content immediately after a deploy or content change, by group, URL, or cache tag.
- Purge by group: Full Site, Specific URLs, Specific Image URLs, Cache Tags, or a rule
- Up to 10 URLs or tags per request
- Path-wide matching for URL purges
- Stats and purge logs reflect every purge
Read purge logs
Customer-visible purge history: who purged what, when, and whether it completed or was rejected.
- Date-filtered purge history with domain scoping
- Actor shows the user or Service API
- Rejected purges stay visible
- Rows available over the API too