Requests
Find a request in Logs and read how Easel handled it, from firewall and cache to function work.
Logs lists the requests your project served. Open one to see the deployment, route, firewall decision, cache result, and function work behind it, along with its log lines and Trace.
Find a request
- Open the project and go to Logs.
- In the filter sidebar, pick a Timeline range and an Environment.
- Narrow the list with filters:
- Route, Request Path, Status Code, and Host
- Request Method
- Cache (
HIT,MISS,BYPASS,STALE) - Contains Console Level (Warning, Error, Fatal)
- Deployment ID
- Select a request row.
Type key:value filters in the search box, then select a row to open its details
The search box also takes key:value filters, such as
requestId:<id>, invocationId:<id>, coldStart:true,
firewallAction:any, or path:"/api". The request ID comes from the
X-Easel-Id response header. From the terminal, run
easel logs --request-id <request-id> --expand. See
easel logs.
To start from a chart instead, note the window where a series moved on Observability, then open Logs for the same window and environment. See Metrics.
Request details
A cached request: the firewall allowed it and the CDN served a HIT
Under the request fields, the panel shows:
- Firewall: the action the firewall took. See Security events.
- Cache: the cache result (CDN) and the framework's ISR cache result when there is one. See Cache observability for how to read them.
- A middleware section and a Function section when compute ran. A request served from the CDN cache, like the one above, has neither.
Function fields
The middleware and Function sections show:
| Field | Meaning |
|---|---|
| Route | The function route that handled the request |
| Compute duration | Time your code ran, through the end of the response body |
| Post-response | Work after the response finished, such as waitUntil |
| Cold start | Whether this wake started a new runtime process |
| Invocation ID | The function wake that served the request |
| Total compute | Middleware and function compute combined |
The duration, billed duration, init, and instance fields describe the whole wake, not just this request. One wake can serve several requests, so open the invocation to see them all. See Function observability.
Logs and Trace
The dock under the request list has two tabs for the selected request:
- Logs: function and platform log lines. See Logs.
- Trace: the span tree for edge and function work. See Tracing.
Build failures aren't part of a request. Read them in the deployment's build logs.