Promhub Logo Prom hub

Host unusual disk read latency
warning

Description Disk latency is growing (read operations > 100ms)
Query for 2m 
>>>
	
				
					(
				
			
				
					
					
						rate
					
				
			
				
					(
				
			
				
					
				
			
				
					[1m]) / 
				
			
				
					
					
						rate
					
				
			
				
					(
				
			
				
					
				
			
				
					[1m]) > 0.1 and 
				
			
				
					
					
						rate
					
				
			
				
					(
				
			
				
					
				
			
				
					[1m]) > 0)
Query Explanation

The rule computes the average read latency per disk by dividing the per‑second increase in node_disk_read_time_seconds_total by the per‑second increase in node_disk_reads_completed_total over the last minute. It triggers when this latency exceeds 0.1 seconds (100 ms) and there is at least one read operation (rate(node_disk_reads_completed_total[1m]) > 0).

Get Alert✕
Download
Copy to Clipboard