You've already forked prometheus-postgres-exporter
							
							doc(README): Avoid deploying on same node / bare metal host
	
		
			
	
		
	
	
		
	
		
			All checks were successful
		
		
	
	
		
			
				
	
				continuous-integration/drone/push Build is passing
				
			
		
		
	
	
				
					
				
			
		
			All checks were successful
		
		
	
	continuous-integration/drone/push Build is passing
				
			This commit is contained in:
		
							
								
								
									
										29
									
								
								README.md
									
									
									
									
									
								
							
							
						
						
									
										29
									
								
								README.md
									
									
									
									
									
								
							| @@ -62,7 +62,7 @@ for customizations. These can be configured in more detail via `values.yaml`. | ||||
|  | ||||
| The following examples serve as individual configurations and as inspiration for how deployment problems can be solved. | ||||
|  | ||||
| ### TLS authentication and encryption | ||||
| #### TLS authentication and encryption | ||||
|  | ||||
| The first example shows how to deploy the metric exporter with TLS encryption. The verification of the custom TLS | ||||
| certification will be skipped by Prometheus. | ||||
| @@ -116,7 +116,7 @@ replaced: | ||||
| +   --set 'prometheus.metrics.serviceMonitor.tlsConfig.keyFile=/etc/prometheus/tls/tls.key' | ||||
| ``` | ||||
|  | ||||
| ### Grafana dashboard | ||||
| #### Grafana dashboard | ||||
|  | ||||
| The helm chart includes Grafana dashboards. These can be deployed as a configMap by activating Grafana integration. It | ||||
| is assumed that the dashboard is consumed by Grafana or a sidecar container itself and that the dashboard is stored in | ||||
| @@ -132,6 +132,31 @@ helm install prometheus-postgres-exporter prometheus-exporters/prometheus-postgr | ||||
|   --set 'grafana.enabled=true' | ||||
| ``` | ||||
|  | ||||
| #### Avoid deploying on same node / bare metal host as PostgresDB | ||||
|  | ||||
| As a best practice, avoid running the postgres-exporter on the same node / bare-metal host as the PostgresDB. This is | ||||
| because if the postgres-exporter is running on the same node and this node fails, Prometheus can send an alert about the | ||||
| failure of the node or that the postgres-exporter cannot be reached. However, it is not possible to react based on the | ||||
| metrics that the postgres-exporter explicitly provides. Depending on the configuration of alerts, this may mean that the | ||||
| corresponding notifications are not sent to the right person or group of people. | ||||
|  | ||||
| The following example prevent the postgres-exporter from running on nodes with a PostgresDB. The PostgresDB nodes has an | ||||
| additional label `database=postgres`. The configuration is carried out in `values.yaml`. | ||||
|  | ||||
| ```yaml | ||||
| deployment: | ||||
|   affinity: | ||||
|     nodeAffinity: | ||||
|       preferredDuringSchedulingIgnoredDuringExecution: | ||||
|       - weight: 100 | ||||
|         preference: | ||||
|           matchExpressions: | ||||
|           - key: database | ||||
|             operator: NotIn | ||||
|             values: | ||||
|             - postgres | ||||
| ``` | ||||
|  | ||||
| ## Parameters | ||||
|  | ||||
| ### Global | ||||
|   | ||||
		Reference in New Issue
	
	Block a user