Suggest Telegraf processor/aggregator plugin approach for json data

Hi I am having json data, below is the sample snippet attached:

{
   "GNBID": "1153",
   "beginTime": "1667535381",
   "counters": {
      "applevel": {
         "OR.ENDCPROCEDURE": {
            "SgnbAddAttemptNeiEnb.neighboreNB.0": "1.00",
            "SgnbAddAttemptNeiEnb.neighboreNB.1": "0.00"
         }
      },
      "cellId=153": {
         "OR.ENDCPROCEDURE": {
            "SgnbAddAttemptCell": "0.00"  
         }
      },
      "cellId=155": {
         "OR.ENDCPROCEDURE": {
            "SgnbAddAttemptCelltwo": "0.00"
         }
      }
   },
   "endTime": "1667535381",
   "swVersion": "6.5.04.072"
}

I want to get below 3 metrics with a tag named cell with values as “applevel”,“153” and “155”, rest of the tags would be same and the fields depends on cell.

Currently I was using - unpivot, starlark, pivot and aggregator plugin combination. However, it is not working fine. Is there any way to do that?

I thought of getting tag names using starlark processor after looping through every fields and generate 3 metrics from 1 metric, but that approach is taking too much time. Any way to optimize that?

Can you please add the metrics you finally expect (maybe in influx-line format) for me to better understand your goal. Furthermore, can you share your current configuration to process this JSON?