Analytics & events
The analytics tools let you look back at how your fleet has behaved over time — charting metric trends, reviewing alert history, and investigating the conditions around a past incident.
Metric explorer
Section titled “Metric explorer”The metric explorer appears on the server detail page and lets you chart any available metric over a configurable time window.
Time ranges: 1 hour, 24 hours, 7 days, 30 days, or a custom date-time range.
Granularity: Choose how the data is rolled up:
| Granularity | Source | Best for |
|---|---|---|
| Raw | ~60-second samples (default collection interval, up to 7 days) | Short-window incident investigation |
| Hourly | Hourly averages (up to 30 days) | Daily trend analysis |
| Daily | Daily averages (up to 365 days) | Long-term capacity planning |
When you select Auto, the server picks the appropriate tier based on the range you chose.
Metric selection: Start with the default preset for the server, or click Explore to browse and add any available metric. Metrics are classified as either server-level (heap, disk, load, threads) or per-channel (received, sent, errors, queued, error rate). Click a metric chip to highlight its line; click again to unfocus.
Saved views: Give a combination of metrics, time range, and granularity a name to recall it instantly. Saved views are per-server and persist across sessions.
Alert events timeline
Section titled “Alert events timeline”Also on the server detail page, the Alert Events timeline shows every alert event recorded for that server, newest first.
Filter by:
- Severity: Critical, Warning, or Info
- State: Firing, Acknowledged, or Resolved
Expand an event row to see the rule snapshot captured at fire time — rule name, type, condition, scope, and evaluation window. This snapshot is frozen when the alert fires so you can understand why it fired even if the rule was later changed or deleted.
If a metric is associated with the event, a View this metric button scrolls the explorer to that metric and focuses the time window around when the event fired.
Data retention
Section titled “Data retention”Monitoring data is stored locally on each Mirth server in an embedded database, and optionally synced to the fleet cloud:
| Tier | Retention |
|---|---|
| Raw samples | 7 days |
| Hourly rollups | 30 days |
| Daily rollups | 365 days |
| Alert events | 90 days |
Rollups are computed on a daily (24-hour) cycle. Older raw data is purged after rollup, so charts automatically switch to a coarser granularity as you zoom out beyond 7 days.