Open in ChatGPTOpen in Claude

Monitor service health

APM rolls every health signal for a service into a single state - shown in the service catalog and on the service itself - so you can tell at a glance which services need attention. The Health tab explains that state: it lists the monitoring policies behind it and lets you tune the thresholds that decide it. The Health tab is available for services.

Use it to:

• Triage fast: see whether a service is critical or warning, and which policy is driving the state.
• Tune health to your service: adjust the warning and critical thresholds so the state reflects real user impact.
• Set an account-wide baseline: save a policy for every service, then override it per service where it needs different thresholds.
• Act on a breach: jump from a policy straight to an alert, its query, the Metric Explorer, or the service's cases.

What you need

• Coralogix Application Performance Monitoring (APM) installed and configured.
• Span Metrics enabled for the service: the latency and error rate policies read span metrics.
• The logs error rate policy evaluates logs correlated to the service, and the cases policy evaluates cases open on it. A policy shows as Unmonitored until its data is available.

Health states

• Critical: at least one policy is breached.
• Warning: at least one policy is at risk.
• Healthy: every policy is within range.
• Unmonitored: no policy evaluates for the service yet: either none is enabled, or none has data to evaluate.

A service's overall state is the worst state across its policies: a single critical policy makes the whole service critical. In the drilldown, the Health policies section header summarizes the tab as a count of how many policies are critical, warning, healthy, and unmonitored.

Access the Health tab

1. In your Coralogix toolbar, select APM.
2. Select a service to open its drilldown, then select the Health tab.

The Health tab is available for services. Databases don't have it.

Health policies

APM evaluates a set of predefined policies against each service. Each policy compares a metric to a warning and a critical threshold over a rolling time window. The Health tab lists four:

Policy	What it measures
Latency	Request latency at the selected aggregation. The row is named for the aggregation in force, such as P95 latency.
Error rate	Share of spans that ended in an error.
Cases	Active cases on the service, by severity (P5-P1).
Logs error rate	Share of log entries emitted at error level.

Note

These are the four policies the Health tab tunes. A service's health badge in the catalog can also reflect its apdex score and SLO compliance - both configured elsewhere, and neither appears as a policy here.

The Health table

The Health tab lists every policy in a table with these columns:

• Name: the policy. The latency row is named for its aggregation: for example, P95 latency.
• Description: what the policy monitors.
• Threshold: the policy's warning and critical thresholds.
• Current value: the policy's live value over the selected time range. The Cases policy shows a count of open cases.
• Status: Healthy, Warning, Critical, or Unmonitored.
• Enabled: a toggle that turns the policy on or off.

Enable or disable a policy

Use the Enabled toggle to turn a policy on or off in place. Turning one off opens a Disable policy? confirmation - This policy will no longer affect Service Health. You can re-enable it at any time. Select Disable policy to confirm, or Discard changes to leave it on. A disabled policy stops affecting health and shows as Unmonitored until you re-enable it.

More actions

Each row's More actions menu offers:

• Edit policy: open the policy editor to change how the policy decides health. See Tune a policy.
• View query: show the query behind the policy.
• Create alert: create a metric alert from the policy's query.
• Open in Metric Explorer: open the policy's metric in the Metric Explorer.
• View cases: open the service's cases. This appears for the Cases policy only.

Which items appear also depends on your permissions - see Permissions.

Tune a policy

Select a policy row (or press Enter), or use More actions then Edit policy, to open the Policy editor - a drawer that changes how a policy decides the service's health. Its toolbar has a breadcrumb, Full screen (and Exit full screen) to expand the drawer, and Close. The editor sets:

• Details: the policy name (read-only) and an editable description.
• Entity filters: the services the policy applies to, shown as Service name = service. The editor opens scoped to the service you're viewing; choose any to make the policy the default for every service. A service's own policy always takes precedence over the default: if you widen a policy to every service while the service keeps its own, the editor flags that the service policy still wins.
• Conditions: the thresholds that decide the state. Set the value at which the status becomes Critical and the value at which it becomes Warning; below both, it stays Healthy. For a latency policy, pick the aggregation (P99, P95, P90, or Avg) and a unit (μs, ms, or s) for each threshold. For the Cases policy, set the case priority (P1-P5) at each level instead. The warning threshold must be lower than the critical threshold; for the Cases policy, the warning priority must be lower than the critical priority.
• Rolling time window: the period the policy evaluates over: 1 min, 5 min, 10 min, or 15 min. The Cases policy has no window.
• Enable policy: a toggle in the editor header that turns the policy on or off.

Select Save policy to apply your changes.

Permissions

Editing health policies requires permission to update the service catalog. Without it:

• Edit policy doesn't appear in the More actions menu.
• The Enabled toggle is disabled, with the tooltip You don't have permission to change health policies.

View cases, Create alert, and Open in Metric Explorer likewise appear only when you have the matching permission - to read cases, update metric alerts, or open the Metric Explorer.

Next steps

Break a service's performance down per transaction in Transactions.