iota-prom-exporter

Prometheus Exporter for IOTA fullnode metrics

Latest Changes:

March 6, 2018

• Changed ZM library to a new promise-based library that bubbles up low level errors
• Minor dashboard changes
• Fixed some bugs and added some minor error handling enhancements

February 19, 2018

• Added a "hasValue" label to Confirmation and Confirm Time metrics to enable comparison of transactions with and without value.
• Updated ZMQ processing to provide more verbose event handling
• Fixed some bugs and added some minor error handling enhancements
• Added a new Dashboard file with improved look and feel and incorporation of the hasValue label.

February 3, 2018

• Removed the stresstest table metrics collector and dashboard row. ZMQ is available on every node and is the recommended approach for tracking those metrics.
• Added confirmation time and confirmation rate tracking. This is accomplished by storing firstSeen and confirmed dates for txs in a LevelDB that resides in a folder /db. A Prometheus histogram is collected on confirmation times and provides bucketing results in the new dashboard. This facilitate other future enhancements but I want to test the LevelDB for a while before adding more to it. There is also a pruning routine added to remove txs older than X days. The config flags for this are in the new config-template.js file.
• Fixed some bugs and added some minor error handling enhancements
• Major changes to the dashboard to include relevenat stats from the Confirmation Time/Rate collection
• Added 3 new API calls: getHistogram() which returns the histogram based on the buckets in the config file. getSeenButNotConfirmed() returns a count of all txs that have a seenDate but no confirmedDate as well as the avg time ago of all of these txs. pruneDB() which prunes the db based on the config values. These are all found at http://localhost:9311/[apiCall]

January 21, 2018

• added a flag to determine if market info should be pulled. If you don't want IOT to BTC/ETH data pulled in, set this flag to false.
• change the reconnect interval for zmq to 20 seconds to give IRI more time to recover.
• added nodejs > 8.0 requirement and updated readme

January 16, 2018

• added zmq metrics for the local server. These are optional and require the configuration of a zmq_url in the config file. If you are upgrading and have zmq enabled on your IRI, add the IP and Port where ZMQ is publishing (default port is 5556).
• added the missing transactions sent from the getNeighbors() api
• added a new dashboard JSON file which includes the latest zmq additions
• cleaned up a few things and added some insurance against crashes

What does it do?

Works with Prometheus and (optionally) Grafana to export instrumentation metrics from an IOTA full node as well as pulls metrics from Bitfinex's webservice API for market data and the IOTA stress table for TPS statistics.

If you run Carriota Field, Dave has a nice exporter for Field metrics

The following images are from my IOTA dashboard. The top looks like this:

The next section contains market pricing data and data from the stresstest node:

Finally, I have a template section for all my neighbors:

Nuriel Shem-Tov has a great installer which deploys, among other things, IOTA IRI, IOTA PM, Grafana, Prometheus and this Exporter. Please visit his docs if your intent is to get the whole stack up and running as easily as possible: IRI-Playbook

IOTA Partners also has a copy+paste installation playbook that is also great. Check out both and decide which one is right for you.

Current Metrics

• totalTransactions
• totalTips
• totalNeighbors
• activeNeighbors
• latestMilestone
• latestSolidSubtangleMilestone
• newTransactions by neighbor
• randomTransactions by neighbor
• allTransactions by neighbor
• invalidTransactions by neighbor
• newTransactions by neighbor
• totalTx
• confirmedTx
• tradePrice by trading pair (IOTUSD, IOTBTC, IOTETH, BTCEUR, IOTEUR)
• tradeVolume by trading pair (IOTUSD, IOTBTC, IOTETH, BTCEUR, IOTEUR)
• zmq transactions seen
• zmq transactions seen with value
• zmq transactions confirmed by milestone
• zmq rstats (the output log line in IRI) toProcess
• zmq rstats toBroadcast
• zmq rstats toRequest
• zmq rstats toReply
• zmq rstats totalTransactions
• avg transaction confirmation time
• avg transaction confirmation rate
• transaction confirmation histogram (buckets)

Dependencies

• Nodejs version > 8.0 (due to async await use)
• Prometheus should be installed. Here is a great guide
• node_exporter for Prometheus gives you system level metrics (instructions included in the above guide)
• Grafana should be installed if you want the sexy dashboards

Installation

Do not run npm install as root as it the zeroMQ and levelDB prebuild dependencies will break the install

git clone https://github.com/crholliday/iota-prom-exporter.git
cd iota-prom-exporter
npm install
node app.js

Rename config-template.js to config.js and update the settings inside (change the port of your IRI installation)

The exporter is configured to run on port 9311 so as to comply with the list of export default ports

Once installed and working you will then need to edit the Prometheus config file - /etc/prometheus/prometheus.yml and add a section like the below:

# iota-prom-exporter section
# scrape_interval is optional as it will default to the default
- job_name: 'iota_exporter'
    scrape_interval: 5s
    static_configs:
      - targets: ['localhost:9311']

I find I need to restart the Prometheus service sudo service prometheus restart after adding an exporter.

Test by navigating to http://localhost:9311/metrics

Grafana

Once the above is done, the metrics will be available to be consumed in a Grafana dashboard.

Docker

The following docker image is available: iota-prom-exporter

make sure to mount your config file, otherwise the image defaults to conifg.example.js

docker run -v your.config.js:/exporter/config.js -p 9311:9311 bambash/iota-prom-exporter

you can also build your own image

docker build . -t iota-prom-exporter
docker run -v your.config.js:/exporter/config.js -p 9311:9311 iota-prom-exporter