Metricbeat 7.8 | aws.rds | cpu intermittently missing from multiple events

[william.shipman]

I'm using metricbeat's AWS module to scrape rds data from cloudwatch for multiple aurora databases. I'm currently seeing events missing cpu data at the same time, across multiple database instances, even while the events themselves are populated.

"_source": {
    "tags": [
      "beats_input_raw_event"
    ],
    "@version": "1",
    "host": {
      "hostname": "v1aptp4mya01"
    },
    "event": {
      "duration": 5775143785,
      "module": "aws",
      "dataset": "aws.rds"
    },
    "ecs": {
      "version": "1.5.0"
    },
    "service": {
      "type": "aws"
    },
    "aws": {
      "rds": {
        "transactions": {
          "blocked": 0,
          "active": 0
        },
        "db_instance": {
          "identifier": "v1aptp4mya01",
          "class": "db.r5.2xlarge",
          "status": "available",
          "arn": "arn:aws:rds:us-east-1:*****************"
        },
        "aurora_replica.lag_max.ms": 20.08300018310547,
        "throughput": {
          "network": 1167980.0395552528,
          "dml": 0.9140919742724659,
          "ddl": 0,
          "commit": 0.8476125579617412,
          "network_receive": 552684.3391115026,
          "select": 892.1371470358491,
          "insert": 0.7811331416510163,
          "update": 0.13295883262144959,
          "delete": 0,
          "network_transmit": 615295.7004437501
        },
        "engine_uptime.sec": 5362978,
        "cache_hit_ratio.buffer": 99.99924804520265,
        "disk_usage": {},
        "latency": {
          "dml": 0.20400000000000001,
          "ddl": 0,
          "commit": 2.227372549019608,
          "select": 1.1854272993163062,
          "insert": 0.193531914893617,
          "update": 0.2655,
          "delete": 0
        },
        "free_local_storage.bytes": 166421417984,
        "freeable_memory.bytes": 6552133632,
        "cache_hit_ratio.result_set": 3.267376587531005,
        "queries": 3078.8389326421425,
        "aurora_bin_log_replica_lag": 0,
        "aurora_replica.lag_min.ms": 20.08300018310547,
        "login_failures": 0,
        "deadlocks": 0,
        "cpu": {
          "total": {}
        },
        "db_instance.identifier": "v1aptp4mya01",
        "database_connections": 148
      },
      "tags": {}
    },
    "metricset": {
      "period": 60000,
      "name": "rds"
    },
    "agent": {
      "version": "7.8.1",
      "hostname": "v1blq2hpmn01",
      "type": "metricbeat",
      "ephemeral_id": "5dc780c0-fa8a-4be4-b15f-3758c2631c00",
      "id": "9bf7e4a8-46f2-4f51-bf0e-a287e299f0e9",
      "name": "v1blq2hpmn01"
    },
    "@timestamp": "2020-11-12T23:41:08.359Z"
  }

I've noticed that this happens at the same time for all db instances, regardless of current activity on the instance.

No errors are reported from metricbeat.

Any input you can provide would be most appreciated. Thank you.