I feel this is true but wanted to confirm. Also, if http_listener is the right input plugins to capture application metrics?
Thanks.
The http_listener input was named long ago and should really have been called influxdb_listener, as it provides the same API for accepting metrics as InfluxDB does. Normally it is used when you want Telegraf to act as a proxy to InfluxDB.
If you want to send application metrics to Telegraf over HTTP, this is the plugin to use, but there are other alternative protocols and formats such as statsd, prometheus, udp/tcp sockets that are also often used for application metrics.
The inputs do generally run concurrently with the outputs, but there is a situation where outputs will pause inputs until they finish writing and you will probably run into this if you try to scale Telegraf up.
