Open the CDN zone → Analytics tab in my.cubepath.com.
NoteAnalytics only become available once the zone is active.
The four headline numbers
| Metric | What it means | What to do about it |
|---|---|---|
| Total Requests | Requests served in the range | Your baseline. A sudden change is either a campaign or an attack |
| Total Bandwidth | Bytes delivered | This is what you're billed for. Cross-check with Pricing |
| Cache Hit Rate | Share of requests answered from the edge | The number that decides your speed and your origin cost. Low hit rate = missing or too-short cache rules |
| Error Rate | Share of 4xx/5xx responses | Rising 5xx usually means the origin is struggling; rising 4xx often means broken links or a WAF rule that's too aggressive |
Charts
- Requests Over Time and Bandwidth Over Time — traffic shape. Read them together: bandwidth climbing while requests stay flat means the files got bigger, not the audience.
- Cache Hit / Miss — where the misses cluster in time. Misses spiking after a deploy usually means your asset filenames changed.
- Status Codes — the health of what you're serving.
- Requests by PoP — which edge locations serve your traffic, which tells you where your audience actually is.
- Blocked Requests — what the WAF is stopping.
Top-N tables
Top URLs, Top Countries, Top User Agents and Top ASN break the traffic down by what, where and who.
These are where the useful surprises live: a single URL taking a disproportionate share of bandwidth is a caching opportunity; a country you don't sell to at the top of the list is usually scraping; and one ASN dominating the traffic is either a large customer or a bot farm.
How to use it
- Right after a cache rule change, watch the cache hit rate. It should move within minutes.
- When a page feels slow, check whether its URL shows up as a cache miss before blaming the origin.
- When the bill surprises you, sort by bandwidth: the answer is almost always one large uncached asset.
- During an incident, put Blocked Requests and Status Codes side by side to tell an attack apart from a broken deploy.
TipTurn on live mode while you're testing a rule change. It's the difference between confirming a fix in thirty seconds and refreshing a dashboard for ten minutes.