The chart provides an operational summary of the DNS seed system as a whole by integrating the data obtained from individual DNS seeds and grouping it into four orthogonal categories.
The duplicate category comprises nodes received from multiple DNS seeds. Ideally, there would be no duplicates but since each DNS seed acts independently, there is a nonzero probability of multiple nodes advertising the same node. Note that if a node is provided by more than one DNS seed, the first occurrence of the node is not included in this category; only the following occurrences are.
The unreachable category covers nodes to which a connection could not be established.
The stale category includes nodes to which a connection could be established but which had already been advertised by the DNS seeds on the day before. Ideally, there would be no overlap between the set of nodes received on successive days. Some potential reasons for overlap include chance, seed(s) using a stale data set, or seed(s) having hard-coded nodes.
The pristine category contains nodes that do not fall into any of the lower-quality categories above. Thus, it includes nodes that were not duplicates, to which a connection could be established, and which had not been advertised by the DNS seeds on the previous day.
A breakdown of DNS seed data by individual seeds is available via the metrics provided in the sidebar and include:
- Number of advertised nodes per seed
- Number and share of pristine nodes per seed
- Number and share of reachable nodes per seed
- Number and share of unreachable nodes per seed
- Number and share of duplicate nodes per seed
- Number and share of stale nodes per seed
To learn about the methodology used to obtain data for the DNS seed metrics, see methodology in the sidebar.The data displayed in the chart is licensed under AGPL and freely available on GitHub.