Log metrics vs "log to business events"

[henk_stobbe]

Hello,

Except for the name, what is the difference. Only a different bucket?

Or in other words when do you use what (-;

KR Henk

[p_devulapalli]

@henk_stobbe 

I guess Log Metrics provides insights into system performance and used for metric analysis and alerting , where as business events are for extracting specific business relevant data from log files and converting it into some meaningful information like certain Id's or costs or customer interactions or business kpi's for business analysis

[t_pawlak · Accepted solution]

What they are

Log metrics → turn log lines into time-series metrics (counts, sums, pXX, etc.) that you chart and alert on via Metric events. Created in Settings → Log Monitoring → Metrics extraction. Stored as metrics powered by Grail.

Log → Business events → convert selected log lines into Business Events (rich, record-level events with attributes like event.type, event.provider, accountId, price), via OpenPipeline. You then query them with DQL, build dashboards/workflows, and (if needed) derive metrics from them. Stored as events in Grail.

So… is it “just a different bucket”?
Not exactly. Both end up in Grail, but as different data types (metrics vs. events) that can live in different buckets and have different retention and usage semantics. (You can manage/create custom buckets per data type.)

When to use what

Use Log metrics when you need simple trends & thresholds fast (e.g., “>100 errors in 5 min”, SLO charts). Lowest friction to alerting.

Use Log → Business events when you need business context and drill-downs (who/what/where), longer retention, ad-hoc DQL, workflow triggers—then optionally create metrics from those events for charts/alerts.

Licensing/retention (high level)

Both are Grail records governed by your Events/Metrics on Grail model (ingest/process/retain). Business Events follow the Events powered by Grail usage model; metrics follow Metrics powered by Grail. Buckets let you tune retention and access per use case.

Rule of thumb

Trend & alert only? → Log metrics.

Business analytics + context (and maybe alerts later)? → Log → Business events (and derive metrics if needed)