# What is a Prometheus target?

In Prometheus, a target is an endpoint or service that Prometheus monitors by scraping metrics. These metrics are exposed in a Prometheus-compatible format, typically over an HTTP or HTTPS endpoint. Targets are configured in the `prometheus.yml` file and identified by two labels: 

- `instance`: Specifies the target’s address (e.g., `192.168.1.101:9090`).  
- `job`: Groups targets logically (e.g., `web-servers`).  

---

[ad-logs-small]

### Configuring targets  

Targets can be defined either statically or dynamically:  

1. **Static configuration**  
   Targets are explicitly listed in the `prometheus.yml` file:  
   ```yaml
   scrape_configs:
     - job_name: 'web-servers'
       static_configs:
         - targets:
             - '192.168.1.101:8080'
             - '192.168.1.102:8080'
   ```

2. **Dynamic service discovery**  
   Prometheus integrates with platforms like Kubernetes, AWS, or Consul to discover targets automatically.  
   Example for Kubernetes:  
   ```yaml
   scrape_configs:
     - job_name: 'kubernetes-pods'
       kubernetes_sd_configs:
         - role: pod
       relabel_configs:
         - source_labels: [__meta_kubernetes_pod_label_app]
           action: keep
           regex: 'web'
   ```

---

### Inspecting targets  

To view the status of targets:  
- Open the Prometheus web UI and navigate to **Status > Targets**.  
- Details include:  
  - State: Whether the target is `UP` or `DOWN`.  
  - Labels: Assigned labels such as `job` and `instance`.  
  - Last Scrape: Timestamp of the last scrape.  
  - Scrape Duration: Time taken for the scrape.  

---

### Troubleshooting  

- If a target is `DOWN`, check its `/metrics` endpoint and verify it exposes metrics in Prometheus format.  
- For missing targets:  
  - Static targets: Confirm the correct IP and port in `prometheus.yml`.  
  - Dynamic targets: Verify service discovery and relabeling configurations.  
- Address label conflicts by using relabeling to modify or filter labels.  

---

### Best practices  

- Use meaningful `job_name` values (e.g., `web-servers` or `db-cluster`) to organize targets logically.  
- Set appropriate scrape intervals to balance resource usage and data freshness:  
  ```yaml
  scrape_interval: 30s
  scrape_timeout: 10s
  ```
- Use relabeling to exclude unnecessary targets and optimize resource usage.  
- Monitor target states and configure alerts for targets that are consistently `DOWN`.  

Targets are the foundation of Prometheus monitoring. Proper configuration and management ensure reliable and efficient metric collection from your systems.

[ad-uptime]