WIP: Telemetry Proposal about 0x-mesh HOT 2 CLOSED

@opaolini does Prometheus support log aggregation? I have experience setting up an ELK stack (Elastisearch, Logstash, Kibana) for log aggregation and it does a lot of the same things that you mentioned here while also being more flexible.

from 0x-mesh.

Prometheus: monitoring, metrics, metrics aggregation and alerts.
Grafana: Nice visualization dashboard for the above
ELK: log transportation via logstash/flutend, aggregation on elastichsearch and kibana for log visualization

Currently, Prometheus is usually deployed alongside an ELK/EFK stack, and I have used it in the past to monitor the ELK stack itself (especially an elasticsearch cluster)

It is true, that you could do some of the things by parsing the event data from logstash and later visualize them on kibana. Also it seems Elastic is getting more into the metrics territory of observability, introduced metricbeat and embraced Prometheus metrics standard

However, ELK stack itself was not designed for time series data and metrics first in mind.

All in all I think it is the question of what is the exact purpose for implementing metrics/observability and how the instrumentation will be used.

As an example if you will be running a local simulation of a mesh network, I don't think you'd want to spin up a an elastichsearch cluster (resource intensive) and then spend processing power parsing logs / events to grab the information like how many orders were rejected, or how many peers are connected. With the prometheus way and pull design, you simply grab that infromation from localhost http://mesh-node-0:3000/metrics , http://localhost:3001/metrics, ..etc without ever needing to spin up prometheus itself.

Also if you will be operating the boostrap nodes on a k8s you will most likely be already running prometheus anyway :)

from 0x-mesh.