Analytics API vs Proxy Dashboard data difference

[rajeevsathyanar]

I am using Analytics API to fetch the sum(message_count) for the proxies for specific date. If i check the data from proxy performance dashboard for same set of proxies and date, i see there is difference in traffic data reported. Proxy dashboard seems to show higher traffic. There is no difference if the traffic is less, but the offset seems to increase considerably for higher traffic. Please let me know how to fix this?

[#2]

hi @Rajeev S, can you paste the actual API call, results and also the proxy performance screenshot highlighting your observation?

[rajeevsathyanar]

Hi @padmanabhans Following is the query

curl -v -X GET "http://MSIP:Port/v1/o/<<orgname>>/environments/<<env>>/stats/apiproxy?select=sum(message_count)&timeRange=06/12/2017%2003:00~06/19/2017%2003:00&timeUnit=week" -u <sysadminid:pwd>

Output for Proxy1

, {
      "metrics" : [ {
        "name" : "sum(message_count)",
        "values" : [ {
          "timestamp" : 1497830400000,
          "value" : "2888544.0"
        }, {
          "timestamp" : 1497225600000,
          "value" : "6.0883559E7"
        } ]
      } ],
      "name" : "Proxy1"
    },

Output for Proxy2

 {
      "metrics" : [ {
        "name" : "sum(message_count)",
        "values" : [ {
          "timestamp" : 1497830400000,
          "value" : "20.0"
        }, {
          "timestamp" : 1497225600000,
          "value" : "977.0"
        } ]
      } ],
      "name" : "Proxy2"
    }

proxy1 numbers are way off(almost by 680k) where as proxy2 numbers are matching.

Please refer attachments for proxy1 and proxy2 Proxy performance screenshot.

NOTE: GUI shows time in PST whereas the server is in EST. Hence the offset(3hrs) in the query.

proxy1.jpgproxy2.jpg

[#4]

Rajeev, thanks for the details.

Can you try your api call with 07:00 as the offset. The api is normally treated as UTC time (not ET) and that might be the reason you are observing the difference for your high traffic Proxy 1

[jonesfloyd]

Hi @Rajeev S - The UI dashboards show metrics at your local time, and the API gets metrics according to UTC. So either change the UI times to match UTC or change the date/time in your API call to match your time in UTC.