Configuring the Pulse Agent
The following are various additional configurations for the Pulse Agent.
Setting the Heap Size
For clusters with a high number of indices and shards, or when monitoring several clusters, running the agent may result in java.lang.OutOfMemoryError
. To address that, set the environement variable PULSE_AGENT_HEAP_SIZE
when running the agent.
The default value is 2G - you'll need to use a larger value, keeping in mind the amount of memory available on the machine hosting the agent. The following one-line docker for running the agent sets it to 8G:
docker run -e PULSE_AGENT_HEAP_SIZE=8G -v `pwd`/config.yml:/etc/pulse-agent/config.yml --restart=always r.bigdataboutique.com/pulse-agent
Controlling Parallelism
By default, the Pulse Agent will use a single thread for effeciency. It will automatically adjust as needed, but you can control the number of threads manually by setting the threads
parameter:
threads: 2
Proxy Configurations
To let the agent use a proxy white sending the metrics to the Pulse platform, add the following settings to the configuration file:
proxyHost: 172.0.0.1
proxyPort: 80
Of course, use the right proxy host and port.
The agent will default to the system proxy settings, if any.
To explicitly completely disable any proxy configurations, use:
proxyHost: disable
Prometheus Endpoint
The agent exposes a Prometheus endpoint for monitoring the Pulse Agent activity, it can be enabled by adding the below to the configurations file:
prometheus:
enabled: true
port: 9090