Metrics
Two metrics extensions implement the MetricsProvider interface on top of Micrometer: Prometheus is pull-based and exposes a scrape endpoint, InfluxDB 2.x is push-based and writes on an interval. Both export the same Universe metrics, so only one can be active at a time.
The metrics extensions surface Universe’s runtime state to your observability stack. Both are built on Micrometer and register the same set of Universe-specific gauges plus the standard JVM and system binders. They differ only in delivery: Prometheus scrapes a REST endpoint, InfluxDB receives pushed line-protocol writes.
Only one MetricsProvider can be active at a time. Install either the Prometheus JAR or the InfluxDB JAR, not both.
Exported metrics
Both extensions record the same Universe-specific metrics:
| Metric | Type | Tags | Meaning |
|---|---|---|---|
universe.instances | Gauge | state=online/offline/stopped | Instance counts by state. |
universe.nodes | Gauge | none | Total connected nodes. |
universe.configs | Gauge | none | Loaded configurations. |
universe.node.ram | Gauge | node_id=... | RAM per node, in MB. |
universe.node.cpu | Gauge | node_id=... | CPU per node, from 0.0 to 1.0. |
Both extensions also auto-register Micrometer’s JVM and system binders, adding memory, GC, thread, CPU, and uptime metrics with no extra configuration.
Choosing a backend
Place extension-metrics-prometheus.jar in ./extensions/. No config file is needed. The Master exposes an unauthenticated scrape endpoint in standard Prometheus exposition format:
GET http://<master-host>:7000/api/metricsPoint Prometheus at it:
scrape_configs:
- job_name: 'universe'
static_configs:
- targets: ['localhost:7000']
metrics_path: '/api/metrics'The endpoint returns plain text:
universe_instances{state="online"} 5.0
universe_nodes 2.0
universe_node_ram_megabytes{node_id="node-1"} 1024.0
jvm_memory_used_bytes{area="heap",id="G1 Eden Space"} 6.7108864E7Place extension-metrics-influxdb.jar in ./extensions/ and create ./extensions/metrics-influxdb/config.json. Universe pushes metrics to InfluxDB 2.x on the configured interval:
{
"url": "http://localhost:8086",
"token": "your-influxdb-token",
"org": "universe",
"bucket": "metrics",
"intervalSeconds": 15
}| Field | Default | Description |
|---|---|---|
| url | "http://localhost:8086" | InfluxDB server URL. |
| token | "" | InfluxDB API token (required). |
| org | "universe" | Organisation name. |
| bucket | "metrics" | Target bucket. |
| intervalSeconds | 15 | Push interval. |
With InfluxDB, metrics are pushed from Universe rather than scraped, so the Master does not need to be reachable from your monitoring host. Only InfluxDB must be reachable from Universe.