Health check
aspic opened this issue · comments
Could it be an idea to have a health-check endpoint? This endpoint could carry out a number of checks to ensure that the Node is behaving properly as for instance:
- RPC call to see that the node is responding
- RPC call to check node values (not sure what is most relevant endpoint/metrics here)
- Let proxy perform a send and listen for receive to make sure that blocks are broadcasted accordingly (advanced)
What are your thoughts?
Maybe. I'm currently using uptimerobot.com and calling https://proxy.nanos.cc/proxy/?action=version and reacting to the expected response. I know if the node or proxy goes down because I also have a similar test for the nodeMonitor software. So personally I'm not in need of such tool.
How would the heath-check notify you?
Uptimerobot would still serve the "notification", but it could check a /health endpoint which could have some more sophisticated tests (configurable of course), in addition to the version-action or other relevant RPC checks.
Ok, then it makes more sense! I'm positive