Project overview

A project opens in Pods on its Overview, the first entry in the sidebar. It sums up traffic across every API in the project over the last 14 days. All its numbers come from live gateway traffic; until a deployed stage serves requests, the charts say there's no traffic yet.

The top right counts the project's APIs, its API Keys and the alarms In Alarm. Beneath, four figures cover the last 14 days, each with its change from the 14 days before:

• Requests: every request the gateway answered.
• Errors: 4XX and 5XX responses together.
• p90 Latency: the time nine in ten requests finished within.
• Throttled: requests turned away by a throttle or quota.

A rise in errors, latency or throttling counts as a change for the worse.

The Overview's charts break the last 14 days of traffic down by day, by API and by outcome.

• Traffic Over Time: a daily line for the metric you pick in Trend metric: Requests (the default), Errors or Throttled. Use the API filter beside it to show all APIs or only the ones you select.
• Traffic by API: each API's share of requests, for the five busiest APIs plus Other. Pick one in Highlighted API to see its number.
• Request Outcomes: the share of requests that Succeeded (2XX and 3XX), got a 4XX or a 5XX, were Throttled or over quota, or were served from the stage cache.
• Availability: the share of requests answered without a 5XX.
• Cache Hit Rate: the share of stage cache lookups that found a cached response.

The Availability and Cache Hit Rate gauges each have their own API filter too.

Top APIs ranks the project's APIs over the last 14 days. Sort it in Sort APIs by: Requests, Error rate or Latency. Each row shows the API's protocol, error rate, requests, p90 latency and health, and the arrow at the end opens the API in the API editor.

• Healthy: under both thresholds for Degraded.
• Degraded: 1% or more of requests got a 5XX, or 10% or more got any error.
• Failing: 5% or more of requests got a 5XX.
• No traffic: no requests in the 14 days.

Needs Attention counts the things worth a look, most urgent first. Click one to go to the screen that deals with it.

• Alarms firing, with the names of the first two, and Awaiting data, alarms with no datapoints yet: both open Monitoring → Alarms.
• Failing APIs and Degraded APIs, by the health rules in Top APIs: open the monitoring dashboard.
• Throttled: requests that hit a rate limit or quota, which opens Usage → Usage Plans.
• Idle APIs: APIs with no traffic in the last 14 days, which opens APIs.