||1 year ago|
|tests||2 years ago|
|zflux||2 years ago|
|.gitignore||2 years ago|
|Dockerfile||2 years ago|
|Dockerfile.build||2 years ago|
|README.md||1 year ago|
|poetry.lock||2 years ago|
|pyproject.toml||2 years ago|
|test-zflux-local.yml||2 years ago|
A buffering proxy for InfluxDB, using a zmq SUB socket.
This was written as a way to improve reliability of InfluxDB, especially when the InfluxDB server lives on a different network than the clients/producers. If the connection between them and the server goes down, or the server is otherwise unreachable, the client has to handle the error and buffer datapoints.
The purpose of
zflux is to sit on the same network as the hosts producing
the data sent to InfluxDB and act as the InfluxDB client. It recieves the
messages from the producers via its SUB socket. If the IfluxDB client cant be
reached, the messages are left in the buffer -- which will be emptied when
the server is reachable.
It supports both using
bind(), but if both are defined it will default to using
The program looks for a config file in the following order
- A full file path defined in the
--- zmq: topic: 'zflux' connect: 'tcp://$zmq_proxy:5560' # alternatively you can use bind() #bind: 'tcp://*:5559' influxdb: host: influxdb.example.com db: $influxdb_database user: $influxdb_username passwd: $influxdb_hostname