Grafana Explore
Grafana Explore is the ad-hoc query interface — the engineer's scratchpad for incident investigation. Unlike dashboards with fixed panels, Explore lets you write any LogQL, PromQL, or TraceQL query against any time range and see results immediately without modifying any dashboard.
Explore Key Features
Single Pane
One query against one data source. Switch between Loki, Prometheus, Tempo without leaving the page.
Split View
Open a second pane side-by-side. Left: Loki logs. Right: Tempo trace. Essential for incident correlation.
Live Streaming
Loki in Explore supports live streaming mode — tails logs in real-time as they arrive. Useful during active incidents.
LogQL Autocomplete
Ctrl+Space shows available labels, label values, and LogQL functions. Dramatically speeds up query writing.
Common Explore Workflows for BizFirstGO
Workflow 1: Find logs for a failing execution
# Select data source: Loki
# Time range: Last 1 hour (or narrow to incident window)
# Query:
{job="processengine", environment="production", level="error"}
|= "executionId=exec-d1e2f3a4"Workflow 2: Find P99 latency for the last hour
# Select data source: Prometheus
# Query:
histogram_quantile(0.99,
sum(rate(bizfirst_node_execution_duration_seconds_bucket[5m])) by (node_type, le)
)Workflow 3: Search for slow traces
# Select data source: Tempo
# Query mode: TraceQL
{ rootName = "workflow.execute" && duration > 10s && span.tenant_id = "tenant-abc" }Explore Split View — Step by Step
Open Explore and run a Loki query
Navigate to Explore. Select Loki. Run a log query to find error logs for an execution.
Click "Split" (the two-pane icon)
A second panel opens on the right, initially showing the same query. Change the right pane's data source to Tempo.
In the left pane, find a log line with a TraceId
Click the TraceId Derived Field link in the log line. The trace opens in the right pane — now you see logs on the left and the corresponding trace on the right simultaneously.
Loki Live Streaming
# To stream logs in real-time in Grafana Explore:
# 1. Select Loki data source
# 2. Enter your LogQL query (must be a log query, not a metric query):
{job="processengine", environment="production"}
# 3. Click the "Live" button (play icon with clock)
# Logs stream in real-time as they arrive in Loki
# Useful for: watching logs during a deployment, monitoring an active incident