Open patoarvizu opened 4 years ago
strongly agreed that it would be helpful if there are configured fields:
Example prefix: staging
staging{"time":"2021-01-22T22:19:10.594844586Z","type":"request","auth":{"token_type":"default"},"request":{"id":"94axxxxxx6","operation":"update","mount_type":"utoken","namespace":{"id":"root"},"path":"auth/xxxxx/login","data":{"jwt":"hmac-sha256:6efb53xxxxxx30f"},"remote_address":"10.15.xxx","headers":{"client-id":["e9fd628xxx"],"client-version":["0.1"],"rpc-caller":["myprober"],"x-request-id":["d8ec2xxxxx"]}}}
prefix value here breaks the JSON blob.
It would make sense that even there is only clusterID there with hmac.
Internally, we can have clusterID map to figure out which cluster on our own.
But, it would be convenient if we can directly add fields.
Is your feature request related to a problem? Please describe. I'm shipping logs from multiple Vault clusters to a central logging facility and I would need to have a way to differentiate where each line comes from. Right now the
prefix
field kinda helps, but because it's not a field in the document, it makes it hard to parse/query.Describe the solution you'd like An additional argument in the audit configuration called something like
additional_attributes
that's a map of fields to be injected into each audit log line.Describe alternatives you've considered Using the configured
prefix
to parse it and inject it, but this solution is kinda fragile.Explain any additional use-cases Potentially, the fields can also be templatized, e.g.