Grafana observability has Grafana for visualization and analytics, Loki for logs, Tempo for traces and Prometheus for metrics while
ELK stack has tools like Elasticsearch for search, Logistash for logs and Kibana for visualization.
Which of these 2 platforms would you choose and why?
AVP, Site Reliability Engineer at a financial services firm with 10,001+ employees
02 October 21
Hi @Faustine Chisasa,
I will go with ELK + Grafana.
Elasticsearch is highly distributed, scalable and fast. You got power of storage. Logstash is highly pluggable and rich in plugins. You can use any buffer or streaming platform in Logstash pipeline example: RabbitMQ, Kafka, redis, etc. ELK has watchers alerts.
S3 searchable snapshot is the best thing in the latest version of ELK which gives you long retention capabilities and search directly from S3 without restore to Elasticsearch.
Grafana can be used to connect to Elasticsearch and you can use all dashboard templates for monitoring.
Metricbeat gives you metrics from On-premise, EC2, Pivotal, cloudfoundry, cloudwatch.
Functionbeat gives you cloudwatch logs.
Grafana needs a separate license and native agents. So all in all, a premium ELK license with open-source Grafana is the best stack for observability.
It's important to do your homework with this solution. I think overall it's a fine product, but you really need to assess what your current needs are, what your future needs are, and see if there's any other competing products out there that might be suitable. My interest at this point is not to replace Elasticsearch. My interest is to either replace or supplement what we have with a different visualization tool. I'd rate this solution an eight out of 10. There's definitely room for improvement.