Device id:
The background reason for this is talaria will be writing to kafka, so we need a consistent partition key, as order matters to the consumers in some use cases, so events for the same CPE should hash to the same partition.
Therefore we always want device id (currently mac address) available in a consistent location on every wrp message. Sometimes it is in the source, sometimes destination, sometimes nowhere. It would be more consistent as a field in metadata set by talaria, if possible.
Account id:
We will be using account id to determine which region to write a message to (different from kafka partitioning, this is more a higher level bucket). All events from the same account must always go to the same region as this is also expected by consumers.
FYI, this is the current metadata from parodus:
firmware string
hardware string
manufacturer string
serialNumber string
lastRebootReason string
protocol string
bootTime string
bootTimeRetryDelay string
interfaceUsed string
Device id:
The background reason for this is talaria will be writing to kafka, so we need a consistent partition key, as order matters to the consumers in some use cases, so events for the same CPE should hash to the same partition.
Therefore we always want device id (currently mac address) available in a consistent location on every wrp message. Sometimes it is in the source, sometimes destination, sometimes nowhere. It would be more consistent as a field in metadata set by talaria, if possible.
Account id:
We will be using account id to determine which region to write a message to (different from kafka partitioning, this is more a higher level bucket). All events from the same account must always go to the same region as this is also expected by consumers.
FYI, this is the current metadata from parodus: