skywalking-webapp shows that the call duration of the trace is incorrect

[snpcp]

The maximum length of the call should be 10007ms, but 90s is instead, so the call length shown is incorrect.

The response data of the trace obtained can be viewed through the network.

This is JSON data:

{
  "data": {
    "trace": {
      "spans": [
        {
          "traceId": "1594174262602.771725797.25057",
          "segmentId": "1594174262602.771725797.25057",
          "spanId": 0,
          "parentSpanId": -1,
          "refs": [],
          "serviceCode": "ATG",
          "startTime": 1594174262602,
          "endTime": 1594174272610,
          "endpointName": "/a/b/c",
          "type": "Entry",
          "peer": "",
          "component": "Nginx",
          "isError": false,
          "layer": "Http",
          "tags": [
            {
              "key": "http.method",
              "value": "GET"
            },
            {
              "key": "http.params",
              "value": "http://10.9.1.103/a/b/c"
            }
          ],
          "logs": []
        },
        {
          "traceId": "1594174262602.771725797.25057",
          "segmentId": "1594174262602.771725797.25057",
          "spanId": 1,
          "parentSpanId": 0,
          "refs": [],
          "serviceCode": "ATG",
          "startTime": 1594174272604,
          "endTime": 1594174272610,
          "endpointName": "/a/b/c",
          "type": "Exit",
          "peer": "server-1",
          "component": "Nginx",
          "isError": false,
          "layer": "Http",
          "tags": [],
          "logs": []
        },
        {
          "traceId": "1594174262602.771725797.25057",
          "segmentId": "1594174174460.512703139.71822",
          "spanId": 0,
          "parentSpanId": -1,
          "refs": [
            {
              "traceId": "1594174262602.771725797.25057",
              "parentSegmentId": "1594174262602.771725797.25057",
              "parentSpanId": 1,
              "type": "CROSS_PROCESS"
            }
          ],
          "serviceCode": "User Service Name 1",
          "startTime": 1594174174460,
          "endTime": 1594174174466,
          "endpointName": "/ingress01",
          "type": "Entry",
          "peer": "",
          "component": "Nginx",
          "isError": false,
          "layer": "Http",
          "tags": [
            {
              "key": "http.method",
              "value": "GET"
            },
            {
              "key": "http.params",
              "value": "http://10.9.1.103/ingress01"
            }
          ],
          "logs": []
        },
        {
          "traceId": "1594174262602.771725797.25057",
          "segmentId": "1594174174460.512703139.71822",
          "spanId": 1,
          "parentSpanId": 0,
          "refs": [],
          "serviceCode": "User Service Name 1",
          "startTime": 1594174174460,
          "endTime": 1594174174466,
          "endpointName": "/ingress01",
          "type": "Exit",
          "peer": "upstream service 1",
          "component": "Nginx",
          "isError": false,
          "layer": "Http",
          "tags": [],
          "logs": []
        }
      ]
    }
  }
}

[snpcp]

[snpcp]

You can see that the call time is 90s instead of 10008ms.

[wu-sheng]

@Fine0830 Could you take a look about this data visualization?

[snpcp]

There is another issue, TotalDuration's statistics seem to be reversed.

[wu-sheng]

What do you mean reversed? Have you checked your local env? Are ATG and User Servica Name deployed on different VMs? Do they have OS clock difference?

[snpcp]

I'll make sure that's okay, you can look at the data That I've provided, and when the data goes into Skywalking, What we should care about is the data itself.

Normally, the first row would be TotalDuration, and the second row would be TotalDuration and SelfDuration, But in fact, the second row says TotalDuration.

[snpcp]

The first line.

The second line.

[wu-sheng]

I found these two timestamp in your span

1594174262602
1594174272610

The difference is 10008ms, what is the issue?

[snpcp]

I slept in the my code for 10 seconds, The reason is that I want to check whether the 90s problem is caused by my duration is too small.

[wu-sheng]

Look these two

1594174262602
1594174174466

You definitely are facing some OS clock differences. All the data get from your post. Please verify it more carefully.

[snpcp]

This is my normal data, you see no more data inversion.

[wu-sheng]

Please don't overload the OSS community with the issue you could check. We just help, rather than service.