v1.0.0

The first release of xrpl-vantage: a Prometheus exporter, alert rules and a Grafana dashboard for what an XRP Ledger node sees of the network and where it stands in it.

Highlights

• Failures are visible. A collector that loses part of its data keeps its failure flag up, every section reports when it last completed a run, and a new alert fires when a section's data goes stale.
• The registry check tells the truth. Timeouts, HTTP errors and answers that are not a validator list count as the registry being down, and the lookup alert fires on all of them.
• Traffic metrics you can graph. xrpl_peer_bytes_sent_total stays monotonic as peers come and go, so rate() works, and concentration is measured on current send rates rather than on how long a connection has been open.
• Correct peer metrics. The disconnect counters now appear, wrong-network peers are counted against the node's own network id, and two unreachable list publishers no longer produce duplicate series.
• Alerts that fit the topology. A validator behind two stock nodes no longer raises a permanent low-peer warning; losing one of its stock nodes does. The list-refresh alert no longer restarts its timer when the failure reason changes.
• Tested in CI. Collector tests against canned RPC answers and behaviour tests for the alert rules, on Python 3.9 and promtool 3.14.

Upgrading from an untagged checkout

A few metrics changed shape: the publisher label is now the publisher's key, peer="other" is gone, xrpl_unl_refresh_ok lost its status label and the peer-count alert now depends on the role label from the scrape config. The changelog has a table of every change and what to use instead.

Get started

git clone --branch v1.0.0 https://github.com/XRPL-Lending-Network/xrpl-vantage.git

Then follow Install in the README. To deploy the nodes themselves, see network-validator-operations.

Team

• @LionelTrichet: exporter, alert rules, tests and release
• @EricTsuii: IPv4-mapped IPv6 peer addresses
• @XRPLLendingNet: link to the operations repository