Closed zerthimon closed 6 years ago
I agree. Currently, I'm sending metrics via UDP to carbon-relay, this relays them to carbon-aggregator via pickle where they are rewritten and forwarded to carbon-cache via pickle :/ Wish it could be simpler
You can use carbon-aggregator-cache
to run aggregation and cache in the same process.
@DanCech I find this section confusing.
Why does the carbon-aggregator-cache.py daemon need the relay-rules.conf file given that carbon-cache.py and carbon-aggregator.py do not? Does it also do relaying like carbon-relay.py does?
Also, which settings does the [aggregator-cache] section support in the carbon.conf file? It has no example section in the carbon.conf.example file. Maybe a union of both [cache] and [aggregator]?
That appears to be a typo, the relay-rules.conf would be used by carbon-aggregator
, not carbon-aggregator-cache
.
I haven't used it myself, but it appears that it will look for an [aggregator-cache]
section that can contain all the settings applicable to the aggregator and the cache.
It'd be most awesome if rewrite rules could be implemented in relay and cache, in the same way they are implemented in aggregator. This would allow for greater flexibility when designing metrics namespaces.