Network Traffic Monitor

This tool was built to analyze AWS network throttling and microbursts. To understand baseline you want to checkout the following resources:

• https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-instance-network-bandwidth.html
• https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/compute-optimized-instances.html#compute-network-performance

Network Traffic Monitor exposes network bytes receive and transmit counters in second granularity.

We store all data in a ring data structure with 300 buckets for each metric.

To query raw data use /raw endpoint and query n with 0 < n < 300 the number of buckets from now you want to get.

To query aggregated values use /metrics endpoint. Query n with 0 < n < 300 the number of buckets we should aggregate. Query baseline to set the value up to we ignore the sum of rx and tx bytes.

# metric value
% curl http://localhost:8080/metrics\?baseline\=10000\&n\=110 | jq .
{
  "above_baseline_count": 3,
  "above_baseline_area_sum": 526611
}


# index metric value
% curl http://localhost:8080/raw\?n\=7 | jq .
{
  "receive_bytes": [
    23737390921,
    23737395660,
    23737682567,
    23737686489,
    23737693129,
    23738061441,
    23738205784
  ],
  "transmit_bytes": [
    10272229724,
    10272238330,
    10272243729,
    10272250883,
    10272257617,
    10272272320,
    10272277800
  ]
}