Serverless App Services and deployments

Analytics

Requests, bandwidth, caching, errors and latency for your deployments, updated hourly.

Cloud → Analytics shows traffic for the whole organization; select a Serverless App Service to see just that one. A Serverless App Service's page also has a traffic summary. Every role can see analytics (analytics:read).

What's measured

Every request the edge serves for your deployment hostnames and custom domains counts:

Metric
RequestsRequests served.
BandwidthBytes sent to visitors.
Cache hit rateShare of cacheable requests served from the edge cache (hits) rather than your deployment (misses).
ErrorsResponses with 4xx and 5xx status codes; server errors are the 5xx ones.
Latencyp50 and p95 of the time the edge took to serve a request.

Top lists show the busiest paths, countries, hosts and deployments for a Serverless App Service, and the busiest Serverless App Services for the organization.

Ranges and freshness

Choose the last 24 hours, 7 days, 30 days or 90 days (by day, from daily rollups kept for 400 days). Numbers are rolled up from the edge's access logs once an hour, a quarter past the hour, and each run recalculates the last three hours so late log deliveries are counted. A range ends at the last complete hour, so a new hour shows up after its rollup rather than as a drop to zero. The page shows when the data was last updated; times are in UTC.

Privacy

The access logs used for analytics record the time, host, path, status, bytes, time taken, cache result, country and deployment of each request. They don't record IP addresses, user agents, cookies or query strings. Raw logs are kept for 90 days, hourly rollups for 35 days and daily rollups for 400 days.

MCP keys with analytics:read can read the same numbers with the project_traffic tool.