MELT (Metrics, Events, Logs, Traces)
MELT stands for Metrics, Events, Logs, and Traces. These are the four core telemetry types that make up modern observability. Together they tell you what is happening inside a system and why.
The four pillars and why they fragment
Each type answers a different question. Metrics tell you the numbers, like request rate and latency. Events mark notable occurrences, like a deploy. Logs give you the detailed text record. Traces follow a single request across services. You usually need all four to debug a real problem.
The trouble is that most stacks store each type in a different tool with a different query language. Metrics in Prometheus, logs in Loki or Elasticsearch, traces in Tempo or Jaeger. Correlating across them during an incident means jumping between systems and stitching results together by hand.
The modern goal is to unify MELT in one place so you can query across all four with one language.
How Arc handles MELT (Metrics, Events, Logs, Traces)
Arc can store metrics, events, logs, and traces together as time-series data, queryable with one SQL surface. Instead of correlating across separate tools by hand, you join across all four in a single query.
Arc is a high-performance columnar database. Open Parquet on storage you own, single Go binary, production-ready in 30 seconds.