Logging is configured under the log configuration property.

log:
  provider: http
  opts:
    url: http://logging-provider.org/
  level: warn

Verbosity

level is the logging verbosity, controlling which events will be logged according to their level, i.e. importance.

It can be silent, info, log, warn or error. The default value is log.

Providers

provider specifies how to send logs.

The following providers are available: http, debug, console and custom.

opts are the options passed to the log provider. It is specific to each provider.

If you want to use several log providers or use several configurations, the log configuration property can be an array of objects instead of a single object.

log:
  - provider: http
    opts:
      url: http://logging-provider.org/
  - provider: debug

HTTP log provider

The http log provider sends logs via HTTP.

Provider options:

• url {string} - URL to send the logs to
• method {string} (default: POST) - HTTP method

log:
  provider: http
  opts:
    url: http://logging-provider.org/
    method: PUT

Console log provider

The console log provider prints logs on the console.

The output is prettified but does not contain performance monitoring nor detailed log information, so this is only meant as a development helper.

It uses the following format:

[EVENT  ] [LEVEL] [HOST ID ] [PROCESS NAME] [PID  ] [TIMESTAMP              ] [PHASE   ] [TIME  ] MESSAGE

For example:

[START  ] [LOG  ] [24cdf75e] [Example API ] [26936] [2017-12-14 12:01:58.846] [STARTUP ] [147ms ] Server is ready

This log provider is always enabled, but can be silenced by using level silent.

log:
  provider: console

Debug log provider

The debug log provider prints logs and performance information on the console. Since the output is not formatted, this is only meant for debugging purpose.

log:
  provider: debug

Custom log provider

When using the custom log provider, logs will be passed as parameters to a custom report function:

• it receives the usual parameters including log, measures and measuresmessage
• it can be async or return a promise
• it can be used to simply handle events instead of logging them

Provider options:

• report {function} - function fired with the log information

log:
  provider: custom
  report:
    $ref: report_log.js

Events

Logs are triggered on the following events:

• start: the server is ready
• stop: the server has exited
• failure: a server-side error occurred
• call: a request has completed, i.e. a response was sent back to the client (whether successful or not)
• message: generic message
• perf: performance monitoring

Those events can be triggered during the following phases of the server:

• startup
• shutdown
• request: each client request
• process: anywhere else, e.g. unhandled exceptions or rejected promises

Log information

The payload of logs is a set of all the currently available parameters, which gives insight about everything in the current context.

Values that might be too big, such as payload, responsedata, data and queryvars, are trimmed: only their id and operationName are kept.

Functions parameters

Besides the usual parameters, the following additional parameters are available during logging:

• log {object}: object containing all the other parameters. Values that might be too big are trimmed. It can be safely serialized.
• event {string}: which event was triggered among start, stop, failure, call, message and perf
• phase {string}: when was the event triggered among startup, shutdown, request and process
• level {string}: event importance among info, log, warn and error
• message {string}: generic message summarizing the event or providing extra information
• protocols {object} - list of protocols being served. Only for start events. Also available as the resolved value of the promise returned by apiServer.run().
  • http {object}: HTTP server information
    • hostname {string}
    • port {string}
• exit {object} - list of servers and databases successfully exited or not. For example, { http: true, mongodb: true, ... }. Only for stop events.
• error {object}: exception object. Only for events failure or request when a client-side or server-side error occurs.
• measures {object[]} - list of performance measurements. Only for perf events:
  • category {string}
  • label {string}: name
  • duration {number} - sum of all measures durations, in milliseconds
  • measures {number[]} - each measure duration, in milliseconds
  • count {number} - number of measures
  • average {number} - average measure duration, in milliseconds
• measuresmessage {string}: console-friendly table with the same information as measures
• duration {number} - time it took for the server to startup, shutdown or handle the client request (depending on the event), in milliseconds. Also available in response's metadata.duration property. Only for perf, start, stop and call events.

Performance monitoring

Logs for the perf event are triggered with information about how long it took for the server to startup, shutdown or handle a client request.

The measures, measuresmessage and duration parameters provide the performance information.