mirror of https://github.com/waku-org/nwaku.git synced 2025-01-29 16:15:54 +00:00

tools: move readmes into their respective folders

Makes it easier to link to just one folder in the repo for documentation
purposes without having to also link the parent folder.

Also added a meta file linking to the tools.
Not sure what the others ones do.

Signed-off-by: Jakub Sokołowski <jakub@status.im>

2022-11-15 16:21:51 +01:00

3.5 KiB

Raw Blame History

networkmonitor

Monitoring tool to run in an existing waku network with the following features:

Keeps discovering new peers using discv5
Tracks advertised capabilities of each node as per stored in the ENR waku field
Attempts to connect to all nodes, tracking which protocols each node supports
Presents grafana-ready metrics showing the state of the network in terms of locations, ips, number discovered peers, number of peers we could connect to, user-agent that each peer contains, content topics and the amount of rx messages in each one.
Metrics are exposed through prometheus metrics but also with a custom rest api, presenting detailed information about each peer. These metrics are exposed via a rest api.

Usage

./build/networkmonitor --help
Usage:

networkmonitor [OPTIONS]...

The following options are available:

 -l, --log-level               Sets the log level [=LogLevel.DEBUG].
 -t, --timeout                 Timeout to consider that the connection failed [=chronos.seconds(10)].
 -b, --bootstrap-node          Bootstrap ENR node. Argument may be repeated. [=@[""]].
 -r, --refresh-interval        How often new peers are discovered and connected to (in minutes) [=10].
     --metrics-server          Enable the metrics server: true|false [=true].
     --metrics-server-address  Listening address of the metrics server. [=ValidIpAddress.init("127.0.0.1")].
     --metrics-server-port     Listening HTTP port of the metrics server. [=8008].
     --metrics-rest-address    Listening address of the metrics rest server. [=127.0.0.1].
     --metrics-rest-port       Listening HTTP port of the metrics rest server. [=8009].

Example

Connect to the network through a given bootstrap node, with default parameters. See metrics section for the data that it exposes.

./build/networkmonitor --log-level=INFO --b="enr:-Nm4QOdTOKZJKTUUZ4O_W932CXIET-M9NamewDnL78P5u9DOGnZlK0JFZ4k0inkfe6iY-0JAaJVovZXc575VV3njeiABgmlkgnY0gmlwhAjS3ueKbXVsdGlhZGRyc7g6ADg2MW5vZGUtMDEuYWMtY24taG9uZ2tvbmctYy53YWt1djIucHJvZC5zdGF0dXNpbS5uZXQGH0DeA4lzZWNwMjU2azGhAo0C-VvfgHiXrxZi3umDiooXMGY9FvYj5_d1Q4EeS7eyg3RjcIJ2X4N1ZHCCIyiFd2FrdTIP"

Metrics

Metrics are divided into two categories:

Prometheus metrics, exposed as i.e. gauges.
Custom metrics, used for unconstrained labels such as peer information or content topics.
- These metrics are not exposed through prometheus because since they are unconstrained, they can end up breaking the backend, as a new datapoint is generated for each one and it can reach up a point where is too much to handle.

### Prometheus Metrics

The following metrics are available. See http://localhost:8008/metrics

peer_type_as_per_enr: Number of peers supporting each capability according the the ENR (Relay, Store, Lightpush, Filter)
peer_type_as_per_protocol: Number of peers supporting each protocol, after a successful connection)
peer_user_agents: List of useragents found in the network and their count

Other relevant metrics reused from nim-eth:

routing_table_nodes: Inherited from nim-eth, number of nodes in the routing table
discovery_message_requests_outgoing_total: Inherited from nim-eth, number of outging discovery requests, useful to know if the node is actiely looking for new peers

### Custom Metrics

The following endpoints are available:

http://localhost:8009/allpeersinfo: json list of all peers with extra information such as ip, location, supported protocols and last connection time.
http://localhost:8009/contenttopics: content topic messages and its message count.

3.5 KiB Raw Blame History

networkmonitor

Usage

Example

Metrics

3.5 KiB

Raw Blame History