Doppler

Doppler is a Cloud Foundry component that gathers logs and metrics from Metron agents, stores them into temporary buffers, and forwards them to third-party syslog drains and other consumers.

Doppler scales horizontally — as load increases, more instances can be deployed to spread the load and decrease memory usage, CPU, etc. per Doppler. Metron agents are notified of the additional instances via updates to etcd.

Architecture Within Loggregator

Logging data passes through the system as protocol-buffers, using Dropsonde.

In a redundant Cloud Foundry setup, Loggregator can be configured to survive zone failures. Log messages from non-affected zones will still make it to the end user. On AWS, availability zones could be used as redundancy zones.

Usage

doppler [--logFile <path to log file>] [--config <path to config file>] \
    [--cpuprofile <path to desired CPU profile>] \
    [--memprofile <path to desired memory profile>]

Flag	Required	Description
--logFile	No, default: STDOUT	The agent log file.
--config	No, default: config/doppler.json	Location of the Doppler configuration JSON file.
--cpuprofile	No, default: no CPU profiling	Write CPU profile to a file.
--memprofile	No, default: no memory profiling	Write memory profile to a file.

Emitting Messages from the other Cloud Foundry components

Cloud Foundry developers can easily add source clients to new CF components that emit messages to Doppler. Currently, there are libraries for Go. For usage information, look at its README.

Editing Manifest Templates

The up-to-date Doppler configuration can be found in the doppler spec file. You can see a list of available configurable properties, their defaults and descriptions in that file.