Looking for Cost-Effective Alternatives to New Relic for Monitoring

I've used Grafana and OTel before, and if your setup includes Spark, check out Dataflint. It simplifies the monitoring of Spark jobs, helps identify slow parts, and ultimately can save you time and money.

Switching from New Relic to a Fluent Bit/OTel/Grafana stack is a smart move to save on costs. Just keep in mind the hidden expenses that come with managing your own storage, like Loki or Prometheus. Handling high cardinality data by yourself can lead to performance issues compared to using a managed service.

I suggest using Vector for log shipping alongside the OTel collector for tracing. I don’t have a strong opinion on storage or querying, but it's worth considering!

You've picked a solid stack! The Fluent Bit + OTel + Grafana combo is popular among many who want to avoid overspending for log insights. Just be aware that Fluent Bit can consume a lot of memory if you buffer too much during outages. Additionally, the OTel collector can easily become a single point of failure unless you set it up properly, which is tricky at first. If you want to simplify things, consider having Fluent Bit send data directly to Loki/Prometheus initially and add OTel later when you need more flexibility. I learned the hard way—spent hours troubleshooting logging issues due to a YAML indentation problem!