Monitoring overview
How traces, logs, and metrics from your services are explored inside CollabOps
Traces, logs, and metrics sent by your services are collected into your CollabOps workspace. You can explore them directly from the Services, Traces, Logs, and Metrics screens, and pin the views you keep coming back to as dashboard panels.
Screens
| Screen | What it shows |
|---|---|
| Services | Every service the workspace collects from, with throughput, error rate, latency, and deployment history per service. |
| Traces | The path a single request took across services. Pick one from search and the detail view lays its spans out over time. |
| Logs | Search over collected logs, narrowed by time range and conditions. |
| Metrics | Browse collected metric names and their labels. |
| Dashboards | Queries pinned as panels on one screen — what this set of documents covers. |
| Settings | Issue and revoke ingestion tokens. |
The three signals
These three are the values a dashboard query picks with signal. Which signal answers best depends on the question.
| signal | What it holds | Questions it answers well |
|---|---|---|
traces | The spans one request went through across services | Where did this request get slow, which span failed |
logs | One line an application wrote | What happened at that moment, what was the full error message |
metrics | Numeric values reported over time, with labels | What is throughput or utilization right now, how is it trending |
Traces are collected in full and the error verdict is made in one place on the server. For request-level questions — slow requests, failed requests — traces are more accurate than logs.
Where dashboards fit
On the exploration screens you rebuild the conditions every time a question comes up. A dashboard does the opposite: it keeps the same question in the same place — service RED (rate, errors, duration), before-and-after a deploy, pod resource usage, anything you look at repeatedly.
One dashboard is one JSON document (a definition). Whether you build it in the editor, import JSON, or commit it to a repository, the same document ends up stored.
Next
| Document | Covers |
|---|---|
| Dashboards | The four ways to create one — panels, sections, variables, built-in templates |
| Definition JSON reference | Every field of the definition document and the validation rules |
| Building dashboards as JSON | Writing a definition by hand, import and export, repository sync |