Summary:
Metalogger is a simple and fast syslog server that can be used within a network to write to multiple destinations and run a large set of processors on This allows for a quick and simple way to add a syslog server in any network. Often times device may not adhere to the proper standards of messages. Metalogger also allows the usage of custom formats by utilizing grok. An example of this can be seen with the CiscoXR format.
Example LogParts:
(format.LogParts) (len=10) {
(string) (len=7) "content": (string) (len=29) "Test UDP syslog message 50000",
(string) (len=8) "priority": (int) 14,
(string) (len=14) "somethingExtra": (string) (len=26) "wow this is something else",
(string) (len=8) "hostname": (string) "",
(string) (len=3) "tag": (string) "",
(string) (len=8) "severity": (int) 6,
(string) (len=6) "client": (string) (len=20) "172.31.255.119:47690",
(string) (len=8) "tls_peer": (string) "",
(string) (len=9) "timestamp": (time.Time) 2022-12-12 00:19:57 +0000 UTC,
(string) (len=8) "facility": (int) 1
}
Processors will simply take in the syslog message and perform processing on it. This can be anything from reverse DNS lookup, to device lookup, interface matching and whatver your use case may be. Adding Processors to the metalogger must satisfy the processor interface:
type Processor interface {
Process(parts format.LogParts) format.LogParts
}
They then can be added to the metalogger by passing them in at instantiation.
s := metalogger.NewMetalogger(
metalogger.WithProcessors([]metalogger.Processor{&TestProcessor{}}),
}
Writers can be added to the system to handle what to do with the messages once they are completed. The writer will just need to follow the writer interface:
type Writer interface {
Write(parts format.LogParts)
}
Writer will write the messages to a destination of your choice. The format of
LogParts is a Map[string]interface{}
so they technically can be marshalled in any
format. It is worthwhile that they should however be batched to the destination.
HealthChecks allows the system a way to perform some type of health check and perform a success or failure action. The healthcheck must satisfy the interface:
type HealthCheck interface {
Check() bool
Success()
Failure()
}
Healthchecks are useful to signal if a destination writer can't be reached or any other checks. One interesting case is the fact that if you have the BGP plugin installed you can stop announcing the anycast IP address of the syslog server if problems arise preventing hosts sending syslog traffic to it.