Monitoring Graph Health

Hello would anybody have any advice on daily testing + sanity checks for a tigergraph instance?

For example I used to check on Neo4j that some nodes/edges were unique. This is not required in TG because they will be unique by definition of the primary key.

I’m wondering what are best practices to flag any errors in the graph early?

@emackie This section has a lot of Sanity Checking types of actions. In the past, I used ELK/DataDog to do the monitoring and log processing. The logs themselves contain a lot of Information that can be filtered on.

You could also set up scripts to do some of the Terminal Commands. Troubleshooting Guide :: TigerGraph Server

1 Like

Thanks but I am looking for something more along the lines of what a data scientist would look for to monitor, e.g., connectivity, size

This would be so I didn’t, e.g., add a bunch of unconnected nodes overnight accidently.

I’m wondering if there are any daily metrics which are best practice.