This is a set of scripts to generate a graph of an NSQ topology, like so:
It has two components:
This script builds up the data backing the topology graph. It does so by communicating with nsqlookupd and then with all the nsqds in the cluster. It is likely that you will want to run it on the same machines hosting nsqlookupd, which will have sufficient access to communicate with the nsqd instances. At the very least, you'll want to run it somewhere in your datacenter.
nsq_data can take a regex as an argument to extract a "host class" from a hostname.
For example, in the default configuration, it will extract the host class "nsqlookup"
from the hostname "nsqlookup01.tinyurl.com". Extracting this host class is essential
to coming up with a readable graph.
nsq_data depends on requests and toolz, which are both pip-installable.
It outputs data in json format, with the 'schema': [{host_class: {topic: {channel: [host_class]}}}]
This script takes the output of nsq_data and uses pygraphviz to generate a DOT graph
of your topology.
It can be loaded as a module and run with a custom function applied to make the labels of the edges in the graph more readable (or to prune away some of them).
It depends on pygraphviz and in turn the graphviz library, which might
make it easier to run on your development machine.
Graphviz can write a variety of image formats; it will infer the desired format from the extension of the provided output path
Putting the pieces together, you may wish to run the components like so:
ssh nsqlookup01 "python /path/to/nsq_data.py" \
| python nsq_graph.py --input - --output $OUTPUTFILE