Open lazam opened 2 years ago
Currently, we only observe many chains in the block-time monitor. The other tools operate in a single-chain mode, and it doesn't make sense to continue working without a connection to a specific node.
In this case, we can use multiple nodes for a single chain to enhance stability. We can cycle through the nodes, connecting to the next one when we get disconnected from the current one. See https://github.com/paritytech/polkadot-introspector/issues/355
The application panicked (crashed) when some of the RPC nodes are unreachable.
Logs:
To reproduce:
Expectation:
Introspector will just log the unreachable RPC nodes and continue on scraping other RPC nodes. RPC nodes being unreachable are common. Other option is to have a metric exposed if a node is reachable or not, something like: