G'day CTinMich,

Welcome to the monastery.

Your current solution seems to be doing a lot of unnecessary processing: slurp entire file; split into records; loop through those records. Depending on the size of the input file, that could be using a lot of memory. Depending on the number of records, shift @records could also be doing a lot of unnecessary processing.

The following (simpler) solution has none of those issues.

#!/usr/bin/env perl

use strict;
use warnings;

{
    local $/ = "\ndirectory";

    while (<DATA>) {
        next if $. == 1;
        /(\S+)\D+(\d\S+)\s+(\/\S+|)[^(]+\(\s*([^)]+)/;
        my ($path, $usage, $threshold) = ($3 || $1, $2, $4);
        print "path=$path; usage=$usage; threshold=$threshold\n";
    }
} 

__DATA__
Type                 Path                           Policy       Snap 
+Usage    
-------------------- ------------------------------ ----------- ----- 
+-------- 
directory            /ifs/home/admin                enforcement    no 
+  32K    
    [hard-threshold] ( 1.0G)
    [hard-threshold-exceeded] (no)
    [container]
    [usage-with-no-overhead] (  32K)
    [usage-with-overhead] ( 103K)
    [usage-inode-count] (4)
directory            /ifs/home/ftp                  enforcement    no 
+  31B    
    [hard-threshold] ( 1.0G)
    [hard-threshold-exceeded] (no)
    [container]
    [usage-with-no-overhead] (  31B)
    [usage-with-overhead] ( 6.0K)
    [usage-inode-count] (3)
directory            /ifs/gpd/data/trufusion/dat... enforcement    no 
+  43G    
/ifs/gpd/data/trufusion/data0003
    [hard-threshold] (  80G)
    [hard-threshold-exceeded] (no)
    [container]
    [usage-with-no-overhead] (  43G)
    [usage-with-overhead] (  95G)
    [usage-inode-count] (606)
[download]

Output:

path=/ifs/home/admin; usage=32K; threshold=1.0G
path=/ifs/home/ftp; usage=31B; threshold=1.0G
path=/ifs/gpd/data/trufusion/data0003; usage=43G; threshold=80G
[download]