shenyu
shenyu copied to clipboard
apache
[Task] Enhancement and optimization of "shenyu_execute_latency_millis_bucket" indicator item of metrics plug-in.
Description
The current situation is:
shenyu_execute_latency_millis_bucket{le="0.005",} 3.0
shenyu_execute_latency_millis_bucket{le="0.01",} 3.0
shenyu_execute_latency_millis_bucket{le="0.025",} 3.0
shenyu_execute_latency_millis_bucket{le="0.05",} 3.0
shenyu_execute_latency_millis_bucket{le="0.075",} 3.0
shenyu_execute_latency_millis_bucket{le="0.1",} 3.0
shenyu_execute_latency_millis_bucket{le="0.25",} 3.0
shenyu_execute_latency_millis_bucket{le="0.5",} 3.0
shenyu_execute_latency_millis_bucket{le="0.75",} 3.0
shenyu_execute_latency_millis_bucket{le="1.0",} 1505.0
shenyu_execute_latency_millis_bucket{le="2.5",} 2351.0
shenyu_execute_latency_millis_bucket{le="5.0",} 3349.0
shenyu_execute_latency_millis_bucket{le="7.5",} 5243.0
shenyu_execute_latency_millis_bucket{le="10.0",} 8050.0
shenyu_execute_latency_millis_bucket{le="+Inf",} 26271.0
Task List
1. Add some time intervals.
Because most request responses are more than 10ms. As shown in the picture above. will add： le="20.0" le="30.0" le="40.0" le="50.0" le="70.0" le="100.0" le="150.0" le="200.0" le="250.0" le="300.0" le="400.0" le="500.0" le="600.0" le="700.0" le="800.0" le="1000.0" le="2000.0" le="3000.0" le="5000.0" le="8000.0" le="10000.0" le="20000.0" le="30000.0"
2. remove some time intervals.
Because it is meaningless to count the delay below 1ms. {le="0.005",} {le="0.01",} {le="0.025",} {le="0.05",} t{le="0.075",} {le="0.1",} {le="0.25",} {le="0.5",} {le="0.75",}
The unit of 0.005 is seconds, which means 5ms.
------------------ Original ------------------ From: lianjunwei @.> Date: Thu,May 12,2022 0:24 AM To: apache/incubator-shenyu @.> Cc: Subscribed @.***> Subject: Re: [apache/incubator-shenyu] [Task] Enhancement and optimization of "shenyu_execute_latency_millis_bucket" indicator item of metrics plug-in. (Issue #3399)
Description
The current situation is: shenyu_execute_latency_millis_bucket{le="0.005",} 3.0 shenyu_execute_latency_millis_bucket{le="0.01",} 3.0 shenyu_execute_latency_millis_bucket{le="0.025",} 3.0 shenyu_execute_latency_millis_bucket{le="0.05",} 3.0 shenyu_execute_latency_millis_bucket{le="0.075",} 3.0 shenyu_execute_latency_millis_bucket{le="0.1",} 3.0 shenyu_execute_latency_millis_bucket{le="0.25",} 3.0 shenyu_execute_latency_millis_bucket{le="0.5",} 3.0 shenyu_execute_latency_millis_bucket{le="0.75",} 3.0 shenyu_execute_latency_millis_bucket{le="1.0",} 1505.0 shenyu_execute_latency_millis_bucket{le="2.5",} 2351.0 shenyu_execute_latency_millis_bucket{le="5.0",} 3349.0 shenyu_execute_latency_millis_bucket{le="7.5",} 5243.0 shenyu_execute_latency_millis_bucket{le="10.0",} 8050.0 shenyu_execute_latency_millis_bucket{le="+Inf",} 26271.0
Task List
- Add some time intervals.
Because most request responses are more than 10ms. As shown in the picture above. will add： le="15.0" le="20.0" le="25.0" le="30.0" le="40.0" le="50.0" le="70.0" le="100.0" le="150.0" le="200.0" le="250.0" le="300.0" le="400.0" le="500.0" le="600.0" le="700.0" le="800.0" le="1000.0" le="2000.0" le="3000.0" le="5000.0" le="8000.0" le="10000.0" le="20000.0" le="30000.0"
- remove some time intervals.
Because it is meaningless to count the delay below 1ms. {le="0.005",} {le="0.01",} {le="0.025",} {le="0.05",} t{le="0.075",} {le="0.1",} {le="0.25",} {le="0.5",} {le="0.75",}
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>
I once thought the unit was seconds. However, according to the time consumption statistics of my printed logs, most requests are less than 500ms. 95% of the requests take less than 1s, and the actual time consumption is inconsistent with the statistics returned by the metric interface. If the unit is seconds, the number falling in the range of Le = "1.0" should be large.