Metrics
Built-in Prometheus metrics and dashboards for your services and databases.
SMLL collects metrics from all your resources using Prometheus. View dashboards and query metrics from the VPC monitoring page.
Available metrics
Service metrics
- CPU usage — percentage of allocated CPU being used
- Memory usage — percentage of allocated memory being used
- Network I/O — inbound and outbound traffic
- HTTP request rate — requests per second (web services)
- HTTP error rate — 4xx and 5xx responses
- HTTP latency — p50, p95, p99 response times
- Container restarts — restart count and OOM kill events
Database metrics
- CPU and memory usage — resource utilization
- Connection count — active and idle connections
- Connection pool usage — PgBouncer pool utilization percentage
- Replication lag — seconds behind primary (replica databases)
- Disk usage — storage consumption
Storage metrics
- Bucket size — total bytes stored per bucket
- Object count — number of objects per bucket
Dashboards
The monitoring page shows pre-built charts for each resource. Select a time range to zoom in on specific periods.
Charts auto-refresh while you're viewing them, giving you a real-time view of your infrastructure.
Time ranges
Select from preset time ranges or define a custom window:
| Range | Resolution |
|---|---|
| Last 15 minutes | 15s intervals |
| Last 1 hour | 1m intervals |
| Last 6 hours | 5m intervals |
| Last 24 hours | 15m intervals |
| Last 7 days | 1h intervals |
| Last 30 days | 6h intervals |
Higher resolution is available for shorter time ranges. Metrics are retained for 30 days.
Metric namespacing
All metrics are scoped to your VPC's Kubernetes namespace. You can only see metrics for resources within your VPC — there is no cross-VPC visibility.
Related
- Alerting — set threshold alerts on any metric
- Uptime Monitoring — HTTP endpoint monitoring
- Logs — correlate metrics with log data for debugging