1. Sources #
What the UI queries.
The principle #
Available in: ✓ Open source ✓ Managed ClickStack
A source describes a ClickHouse table and how to read it: time column, message body, level, service name, trace IDs. ClickStack creates default sources with the first account; you can add more for any ClickHouse table, including your own data.
| Source | Table | Use |
|---|---|---|
| Logs | otel_logs | search, live tail, charts |
| Traces | otel_traces | span search, waterfall, latency |
| Metrics | otel_metrics_gauge, _sum, _histogram… | charts and dashboards |
| Sessions | hyperdx_sessions | browser session replay |
Best practice : Link sources together (logs ↔ traces ↔ sessions) in their settings: that is what enables correlation links in the UI.
2. Searching #
Properties, full text or SQL.
Search syntax #
| Search | Finds |
|---|---|
timeout | events whose body contains “timeout” |
"connection refused" | the exact phrase |
level:err | error level (property) |
service:checkout level:err | both conditions (implicit AND) |
service:checkout OR service:payments | either one |
-level:info / NOT level:info | exclude |
http.status_code:500 | an attribute (Map column key) |
The SQL mode accepts a WHERE condition directly: SeverityText = 'error' AND LogAttributes['http.route'] LIKE '/api/%'. Handy for ClickHouse functions (regular expressions, numeric comparisons).
Exploring results #
- Live tail: new events stream in during a release or a test.
- Histogram of event counts over time: click and drag to zoom into a spike.
- Columns: add an attribute as a column, filter on a value in one click from an event's details.
- Saved searches: reusable, shareable, and the basis of alerts.
3. Traces, logs and sessions #
Correlating signals.
The journey #
- Trace: the waterfall shows each span, its duration and attributes; logs emitted during each span are shown alongside.
- From log to trace: a log carrying a
TraceIdopens the full request trace. - From trace to session: if the browser SDK is in place, the user's session, journey and replay.
A typical investigation #
- An alert reports errors: search
service:checkout level:errover the last hour. - The histogram shows when it started; a log's details give the faulty
http.routeattribute. - Open the trace: the longest span is a database call timing out.
- The session shows the user's action and the error message displayed.
- Measure the scope in SQL (number of users affected) before communicating.
4. Going further with SQL #
When the UI is not enough.
Useful queries #
-- Distinct users affected by the error over the last hour
SELECT uniq(LogAttributes['user.id']) AS users
FROM otel_logs
WHERE ServiceName = 'checkout' AND SeverityText ILIKE 'err%' AND Timestamp > now() - INTERVAL 1 HOUR;
-- Slowest routes (p95) from traces
SELECT SpanAttributes['http.route'] AS route, quantile(0.95)(Duration) / 1e6 AS p95_ms, count() AS calls
FROM otel_traces
WHERE ServiceName = 'checkout' AND SpanKind = 'Server' AND Timestamp > now() - INTERVAL 1 HOUR
GROUP BY route ORDER BY p95_ms DESC LIMIT 10;Good to know : In Managed ClickStack, AI-assisted notebooks help write these analyses; always check the generated query before trusting the result.
Commands to know #
Commands to know
service:checkout level:err: property search"connection refused": exact phrase-level:info: exclusionLogAttributes['http.route']: read an attribute in SQLquantile(0.95)(Duration) / 1e6: p95 latency in milliseconds
Sources #
Checked in November 2026: ClickStack evolves fast (Managed ClickStack is in beta), check the ClickHouse documentation before relying on a detail.
My progress
Sign in to track your progress and take notes. Sign in
💬 Questions about this lesson
No questions yet: be the first to ask for help or share your configuration.
Recommended Articles
Exploring data in OpenObserve: SQL, PromQL and VRL
Search with SQL, query metrics with PromQL, optimize a stream (indexes, partitions, retention) an...
Installing and deploying ClickStack: all-in-one, Docker Compose, Helm and Managed ClickStack
Deploying ClickStack to fit the need: all-in-one container with persistence, Docker Compose and i...
Profiles, OpenTelemetry's fourth signal: continuous profiling and correlation with traces
Understand OpenTelemetry's new profile signal: continuous profiling, flame graphs, data model and...
Logs with OpenTelemetry: data model, appenders, correlation and collection
Master the log signal: the LogRecord data model, the Logs Bridge API and appenders (Logback, Log4...
OpenTelemetry traces in depth: spans, status, events, links and sampling
Master the trace signal: anatomy of a distributed trace, span and SpanKind, attributes, status an...
DQL in practice: exploring logs and events, notebooks, dashboards and performance
Use the Dynatrace Query Language day to day: explore logs and events, chain commands, parse field...
Correlating traces, metrics, logs and profiles with OpenTelemetry
Link the four signals to diagnose quickly: shared resource, trace_id in logs, exemplars on metric...
The observability trilemma: cost, completeness and performance, and how solutions resolve it
Understanding the observability trilemma (cost, data completeness, query performance), its three ...
Using SigNoz: APM, traces, logs, dashboards and alerts
Use SigNoz day to day: APM view of services, traces and logs explorers, query builder, ClickHouse...