new GoodInflux(host, [options])
host- The host, including protocol, port, and optional credentials, where InfluxDB is running
database- The name of the database data will be written to. Defaults to
eventsconfiguration property as supported by
threshold- The number of events to queue before flushing to the database. Defaults to 25.
username- The username for the InfluxDB user. This value is overridden by host credentials, if available. (Requires
password- The password for the InfluxDB user. This value is overridden by host credentials, if available. (Requires
precision- The precision specified for InfluxDB timestamp. Defaults to
ms- milliseconds (default)
Each request is recorded as a point in the
request data series. The request series can then be broken down into sub-series via continuous queries. For example, to get requests per process per host you can run
select * from "request" into request.[host].[pid]
Once that query is complete, InfluxDB will dynamically create series for each new
pid combination recorded.
Responses that result in 500 status code are logged as errors. Each error contains the error message and stacktrace, in addition to the url and HTTP method pertaiing to the failed request.
OS-level details are also recorded, including the columns above. NOTE: Currently there is no deduping of data, therefore if multiple processes are running on the same host, duplicate OS data will be recorded to Influx.
process-leve details, including the event loop delay, are recorded into a
Server logs, including tags are also recorded as points in the
When logging a custom event a series will be created using that event name. Additionally, custom events wil automatically get logged with
pid columns. Additionally, any other fields/properties are logged as columns into InfluxDB. Some types are handled specially, however: primitives (string, boolean, number) are logged as-is, dates are converted to date strings, arrays are logged as comma-separated values, and objects are JSON stringified prior to logging.