Read a service's detail page
Go through an APM service's tabs to see its performance, endpoints, errors, logs, dependencies and incidents, and link it to the inventory.
On this page
The detail page brings together everything APM knows about a service. You open it from APM → Services, from the dependency map or from any link to the service.
Before you start
View permission in the APM module. To change the link or the thresholds, Operate permission.
Steps
Go to APM → Services and click the service's name.
Choose the period in the selector on the right. APM shows up to the last 14 days.
Go through the tabs.
The tabs
Tab | What it contains |
|---|---|
Summary | Requests, errors, p50, p95 (with the value from 7 days ago) and p99 for the period; availability (requests without errors); Where it runs (platform, type, language, first and last seen, deployed versions and instances); the Inventory card and the Cerebro's failure risk, if your plan includes it |
Monitors | The telemetry and tests associated with the service |
Charts | Requests and errors, Latency (p50, p95 and p99) and Application metrics: choose a metric from the list to see its chart |
Endpoints and traces | Each endpoint with its traffic, errors and latency. Click one to see its latency chart and its traces; filter by errors |
Errors and logs | The period's errors with their status, and the warning and error Recent logs, with a link to their trace |
Dependencies | Depends on (databases, HTTP, queues, RPC and other services) and Called by, each with its traffic, errors and p95 |
Events | What happened with the service during the period |
Incidents | The service's incidents |
Relationships | Its relationships in the inventory |
Comments | Team notes about the service |
If the Charts tab says the service sends no metrics of its own, check that OTEL_METRICS_EXPORTER=otlp is set in the application.
Link the service to the inventory
A service without a group does not appear in the Inventory or in group dashboards, and its incidents have no group. ARGUNIS tries to link it on its own; if it can't, you do it on the Inventory card (in Summary or in Configure → General).
Choose an Own group, or link it to a Server, Web or TCP service (a
wss://or a custom port is monitored as TCP) or Cloud service (from Railway, AWS, Azure or Google Cloud) that already has a group: the service inherits it.If there are Suggestions, click Link on the right one. Those marked safe match without a doubt.
To undo what you set manually, click Back to automatic. To unlink it completely, Remove link.
What you set manually is marked set manually and ARGUNIS does not change it.
Service thresholds and muting
Each service uses the company's alert thresholds, unless you give it its own.
On the detail page, click Configure and open the Thresholds tab.
Click Edit thresholds and adjust Maximum p95 (ms), Maximum errors (%), Minimum requests in the window, Window (min), Notify if it sends no data for (min) and Close errors that don't recur (days) (from 1 to 90: after that many days without the error returning, its incident closes on its own). Click Save.
To go back to the company's thresholds, click Use the company's.
To keep the service from opening incidents for a while (for example, during a migration), click Mute alerts. Turn them back on with Resume alerts.
Related
What you change on the service's Relationships tab also changes the Dependency map and Live APM: a dependency you hide is no longer drawn, and one you add manually appears dotted (no traffic).