Having different measurement fields in time-series collection documents

Hi all.
I need to store different metrics of our IOT application nodes (log messages, CPU temperature, etc) in a timeseries collection.
Each log record has its own type and each type has its own measurement fields:

{
  "timestamp": ...,
  "info": {
    "type": 0,
    "nodeId": 14
  },
  "temperature": 53
}

{
  "timestamp": ...,
  "info": {
    "type": 1,
    "nodeId": 14
  },
  "log": "Wifi connection failure"
}

{
  "timestamp": ...,
  "info": {
    "type": 2,
    "nodeId": 14
  },
  "inputPowerOutage": true
}

Are there any disadvantages to having different measurement fields in the same timeseries collection (temperature, log, inputPowerOutage)? I should also mention that each type has its own measurement fields, So I guess there aren’t two different type of log records in the same bucket.