Get the size and object count of an S3 bucket

bsize

#!/usr/bin/env bash
#
# show S3 bucket size/count
# usage: bsize <bucket>
#
# author: Erhhung Yuan <erhhung@alum.mit.edu>

bucket=$1

if [ -z "$bucket" ]; then
  echo 'Show S3 bucket size/count'
  echo 'Usage: bsize <bucket>'
  exit
fi

# get first command
# available in $PATH
# _altcmd gsort sort
_altcmd() {
  for bin in "$@"; do
    if hash $bin 2> /dev/null; then
      printf $_ && return
    fi
  done
  return 1
}
numfmt=$(_altcmd gnumfmt numfmt)

# require given commands
# to be $PATH accessible
# _require foo bar || return 1
_require() {
  for bin in "$@"; do
    if ! hash $bin 2> /dev/null; then
      echo >&2 "Please install \"$_\" first!"
      return 1
    fi
  done
}
_require aws jq ${numfmt:-numfmt} || exit 1

# get bucket region
region=$(aws s3api get-bucket-location \
            --bucket $bucket 2> /dev/null | \
  jq -r '.LocationConstraint // "us-east-1"')
if [ -z "$region" ]; then
  echo >&2 "Cannot determine bucket location!"
  exit 1
fi

# https://docs.aws.amazon.com/AmazonS3/latest/userguide/metrics-dimensions.html
stypes=(
  StandardStorage
  IntelligentTieringFAStorage
  IntelligentTieringIAStorage
  IntelligentTieringAAStorage
  IntelligentTieringAIAStorage
  IntelligentTieringDAAStorage
  StandardIAStorage
  StandardIASizeOverhead
  StandardIAObjectOverhead
  OneZoneIAStorage
  OneZoneIASizeOverhead
  ReducedRedundancyStorage
  GlacierInstantRetrievalStorage
  GlacierStorage
  GlacierStagingStorage
  GlacierObjectOverhead
  GlacierS3ObjectOverhead
  DeepArchiveStorage
  DeepArchiveObjectOverhead
  DeepArchiveS3ObjectOverhead
  DeepArchiveStagingStorage)

# _bsize <metric> <stype>
_bsize() {
  utnow=$(date +%s)
  period=$((60*60*24*2))
  metric=$1 stype=$2

  aws cloudwatch get-metric-statistics  \
    --start-time  $(($utnow - $period)) \
    --end-time    $utnow  \
    --period      $period \
    --region      $region \
    --namespace   AWS/S3  \
    --metric-name $metric \
    --dimensions  Name=BucketName,Value=$bucket \
                  Name=StorageType,Value=$stype \
    --statistics  Average 2> /dev/null | \
    jq -r '.Datapoints[].Average // 0'
}

total=$(
  (for stype in ${stypes[@]}; do
      _bsize BucketSizeBytes $stype
    done; echo 0) | \
    paste -sd+ - | bc
)
count=$(_bsize NumberOfObjects AllStorageTypes)

# _print <label> <number> <units> <format> [suffix]
_print() {
  read label number units format suffix <<< "$@"
  echo "$label"

  numfmt $number \
    --to="$units" \
    --suffix="$suffix" \
    --format="$format" | \
    sed -En 's/([^0-9]+)$/ \1/p'
}

cols=($(
  _print Size "0${total}" iec-i "%.2f" B
  [ "0${count}" -lt 1000 ] && echo Count $count || \
    _print Count "0${count}" si "%.2f"
))
printf "%5s: %6s %s\n" "${cols[@]}"