Logs
The Logs page is where you search, cluster, and watch your log lines as they arrive. Search runs server-side, so it stays fast even over large volumes.
Searching
Type into the search box using a lightweight meta-language that mixes free-text keywords with field filters:
- Free text matches against the log message.
project:<namespace>scopes results to a single Project (the namespace). This is the same Project filter you'll see across the dashboard.trace_id:<id>andspan_id:<id>narrow to the logs from a specific trace or span. These are the targets of the "View logs for this trace/span" deep-links on the Traces page, so jumping from a request to its logs lands here pre-filtered.
Results load newest-first, and scrolling down pages in older results automatically (infinite scroll). You can refresh manually or set an auto-refresh interval.
Reading a log line
Click any line to open a side panel with the full record, split into clear sections:
- Attributes — the fields on the log record itself.
- Resource attributes — fields describing the emitter (the service or host that produced it).
- Trace — when the line carries a
trace_id, a link jumps to the originating trace on the Traces page.
Columns
Tune the table to your needs:
- Column picker to choose which fields are shown (including an optional
trace_idcolumn). - Resizable columns and a wrap toggle for long messages.
Patterns
The Patterns tab groups similar log lines into templates automatically (drain clustering). Instead of scrolling thousands of near-identical lines, you see the handful of underlying shapes and how often each occurs — the fastest way to find what's actually noisy or new.
Live Tail
The Live Tail tab streams logs in real time over a WebSocket, like tail -f for your whole environment. Use it to watch a deploy or reproduce an issue as it happens.
Related
- Traces — jump between a request and its logs in both directions.
- Query Explorer — query logs with OQL for charts and aggregations.