|
Line 0
Link Here
|
|
|
1 |
package Koha::Middleware::Throttle; |
| 2 |
|
| 3 |
# This file is part of Koha. |
| 4 |
# |
| 5 |
# Koha is free software; you can redistribute it and/or modify it |
| 6 |
# under the terms of the GNU General Public License as published by |
| 7 |
# the Free Software Foundation; either version 3 of the License, or |
| 8 |
# (at your option) any later version. |
| 9 |
# |
| 10 |
# Koha is distributed in the hope that it will be useful, but |
| 11 |
# WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 |
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 13 |
# GNU General Public License for more details. |
| 14 |
# |
| 15 |
# You should have received a copy of the GNU General Public License |
| 16 |
# along with Koha; if not, see <http://www.gnu.org/licenses>. |
| 17 |
|
| 18 |
use Modern::Perl; |
| 19 |
|
| 20 |
use parent qw(Plack::Middleware); |
| 21 |
use Plack::Response; |
| 22 |
use Plack::Util::Accessor qw(debug interface allow_list paths request_threshold last_n_minutes cidr_block); |
| 23 |
use Net::Netmask; |
| 24 |
|
| 25 |
use C4::Templates; |
| 26 |
use Koha::Caches; |
| 27 |
|
| 28 |
=head1 Functions |
| 29 |
|
| 30 |
=head2 prepare_app |
| 31 |
|
| 32 |
Called at startup time |
| 33 |
|
| 34 |
=cut |
| 35 |
|
| 36 |
sub prepare_app { |
| 37 |
my $self = shift; |
| 38 |
|
| 39 |
if ( !$self->request_threshold ) { |
| 40 |
|
| 41 |
#NOTE: Give a high fallback request threshold |
| 42 |
$self->request_threshold(50); |
| 43 |
} |
| 44 |
if ( !( $self->last_n_minutes && $self->last_n_minutes > 1 ) ) { |
| 45 |
|
| 46 |
#NOTE: Give a reasonable fallback number of last N minutes to check for visits |
| 47 |
$self->last_n_minutes(5); |
| 48 |
} |
| 49 |
my $allowed_cidr_blocks = { |
| 50 |
8 => 1, |
| 51 |
16 => 1, |
| 52 |
24 => 1, |
| 53 |
32 => 1, |
| 54 |
}; |
| 55 |
if ( $self->cidr_block ) { |
| 56 |
if ( !$allowed_cidr_blocks->{ $self->cidr_block } ) { |
| 57 |
$self->cidr_block(24); |
| 58 |
} |
| 59 |
} else { |
| 60 |
$self->cidr_block(24); |
| 61 |
} |
| 62 |
} |
| 63 |
|
| 64 |
=head2 call |
| 65 |
|
| 66 |
Called at request time |
| 67 |
|
| 68 |
=cut |
| 69 |
|
| 70 |
sub call { |
| 71 |
my ( $self, $env ) = @_; |
| 72 |
|
| 73 |
#NOTE: Like the CSRF middleware, we only apply this once in the request cycle. |
| 74 |
#NOTE: We don't apply it to errordocument subrequests. |
| 75 |
if ( !$env->{'plack.middleware.Koha.Throttle'} ) { |
| 76 |
|
| 77 |
my $ip_address = $env->{REMOTE_ADDR}; |
| 78 |
|
| 79 |
my $throttle_check = 0; |
| 80 |
my $threshold = $self->request_threshold; |
| 81 |
|
| 82 |
my $req = Plack::Request->new($env); |
| 83 |
my $request_path = $req->path // q{}; |
| 84 |
my $user_agent = $req->user_agent // q{-}; |
| 85 |
|
| 86 |
my $paths = $self->paths(); |
| 87 |
if ( $paths && ref $paths && ref $paths eq 'ARRAY' ) { |
| 88 |
foreach my $path (@$paths) { |
| 89 |
|
| 90 |
#NOTE: Since we're using flexible regex, we have to iterate through a list rather than using a more efficient hash lookup |
| 91 |
if ( $request_path =~ qr{$path} ) { |
| 92 |
$throttle_check = 1; |
| 93 |
last; |
| 94 |
} |
| 95 |
} |
| 96 |
} |
| 97 |
|
| 98 |
if ($throttle_check) { |
| 99 |
|
| 100 |
#NOTE: If this request requires a throttle check, we first check out allow_list for exceptions |
| 101 |
if ( $self->allow_list |
| 102 |
&& ref $self->allow_list |
| 103 |
&& ref $self->allow_list eq 'ARRAY' ) |
| 104 |
{ |
| 105 |
foreach my $ip_range ( @{ $self->allow_list } ) { |
| 106 |
if ( $self->debug ) { |
| 107 |
warn "Checking IP address '$ip_address' against IP range '$ip_range'\n"; |
| 108 |
} |
| 109 |
my $mask = Net::Netmask->new2($ip_range); |
| 110 |
if ($mask) { |
| 111 |
if ( $mask->match($ip_address) ) { |
| 112 |
if ( $self->debug ) { |
| 113 |
warn "IP address '$ip_address' matched IP range '$ip_range'. Skip.\n"; |
| 114 |
} |
| 115 |
$throttle_check = 0; |
| 116 |
last; |
| 117 |
} |
| 118 |
} |
| 119 |
} |
| 120 |
} |
| 121 |
} |
| 122 |
|
| 123 |
if ($throttle_check) { |
| 124 |
my $total_visits = $self->_get_total_visits( |
| 125 |
{ |
| 126 |
ip_address => $ip_address, |
| 127 |
} |
| 128 |
); |
| 129 |
if ( $self->debug ) { |
| 130 |
warn "Total visits during time period: $total_visits\n"; |
| 131 |
warn "Visit threshold: $threshold\n"; |
| 132 |
} |
| 133 |
if ( $total_visits > $threshold ) { |
| 134 |
my $error = sprintf( "Throttled! IPADDRESS{{%s}} USERAGENT{{%s}}\n", $ip_address, $user_agent ); |
| 135 |
warn $error; |
| 136 |
my $output; |
| 137 |
my $template = $self->_prepare_template( { template_filename => 'opac-throttled.tt' } ); |
| 138 |
if ($template) { |
| 139 |
$output = $template->output(); |
| 140 |
} |
| 141 |
my $res = Plack::Response->new( 200, [ 'Content-Type' => 'text/plain' ], ["Throttled!"] ); |
| 142 |
if ($output) { |
| 143 |
$res->content_type('text/html'); |
| 144 |
$res->body($output); |
| 145 |
} |
| 146 |
return $res->finalize; |
| 147 |
} |
| 148 |
} |
| 149 |
|
| 150 |
} else { |
| 151 |
$env->{'plack.middleware.Koha.Throttle'} = 'checked'; |
| 152 |
} |
| 153 |
return $self->app->($env); |
| 154 |
} |
| 155 |
|
| 156 |
sub _get_total_visits { |
| 157 |
my ( $self, $args ) = @_; |
| 158 |
my $total_visits = 0; |
| 159 |
my $ip_address = $args->{ip_address}; |
| 160 |
|
| 161 |
#NOTE: `memcdump --servers=memcached | grep "throttle"` should show you all the throttle cache keys but it's not... |
| 162 |
my $cache = Koha::Caches->get_instance('throttle'); |
| 163 |
if ( $ip_address && $cache ) { |
| 164 |
my $cache_obj = $cache->{memcached_cache}; |
| 165 |
if ( $cache_obj && ref $cache_obj && ref $cache_obj eq 'Cache::Memcached::Fast::Safe' ) { |
| 166 |
my $cidr = sprintf( "%s/%d", $ip_address, $self->cidr_block ); |
| 167 |
my ($ip_range) = Net::CIDR::cidr2octets($cidr); |
| 168 |
|
| 169 |
#my $ip_range = join('.',(split /\./, $ip_address)[0..2]); |
| 170 |
my $cache_keys = $self->_generate_cache_keys( |
| 171 |
{ |
| 172 |
ip_range => $ip_range, |
| 173 |
} |
| 174 |
); |
| 175 |
my $current_cache_key = $cache_keys->[0]; |
| 176 |
my $get = $cache_obj->get($current_cache_key); |
| 177 |
if ( !$get ) { |
| 178 |
|
| 179 |
#FIXME: Improve expiry time configuration |
| 180 |
#FIXME: Should this time be relative to the time of the cache key's minute-truncated time rather based from now? |
| 181 |
my $expiry_secs = $self->last_n_minutes * 60; |
| 182 |
if ( $self->debug ) { |
| 183 |
warn sprintf( |
| 184 |
"Cache key %s expires in %s seconds (%s minutes)\n", $current_cache_key, $expiry_secs, |
| 185 |
$self->last_n_minutes |
| 186 |
); |
| 187 |
} |
| 188 |
$cache_obj->set( $current_cache_key, 0, $expiry_secs ); |
| 189 |
} |
| 190 |
my $rv = $cache_obj->incr($current_cache_key); |
| 191 |
my $results = $cache_obj->get_multi(@$cache_keys); |
| 192 |
my @sorted_keys = sort { $b cmp $a } keys %$results; |
| 193 |
foreach my $key (@sorted_keys) { |
| 194 |
my $value = $results->{$key}; |
| 195 |
if ($value) { |
| 196 |
if ( $self->debug ) { |
| 197 |
my ( $key_ip_range, $key_time ) = split( ':', $key ); |
| 198 |
warn sprintf( |
| 199 |
"IP range '%s' Time '%s' -> %s visits\n", $key_ip_range, |
| 200 |
scalar localtime($key_time), $value |
| 201 |
); |
| 202 |
} |
| 203 |
$total_visits += int($value); |
| 204 |
} |
| 205 |
} |
| 206 |
} |
| 207 |
} |
| 208 |
return $total_visits; |
| 209 |
} |
| 210 |
|
| 211 |
sub _get_truncated_timestamp { |
| 212 |
my ( $self, $args ) = @_; |
| 213 |
my $time = time; # Get the current time as a Unix timestamp |
| 214 |
$time -= $time % 60; # Truncate by minute interval |
| 215 |
return $time; |
| 216 |
} |
| 217 |
|
| 218 |
sub _generate_cache_keys { |
| 219 |
my ( $self, $args ) = @_; |
| 220 |
my $ip_range = $args->{ip_range}; |
| 221 |
my @cache_keys = (); |
| 222 |
if ($ip_range) { |
| 223 |
my $now = $self->_get_truncated_timestamp(); |
| 224 |
if ($now) { |
| 225 |
|
| 226 |
#Default to last 5 minutes |
| 227 |
my $last_n_minutes = $self->last_n_minutes; |
| 228 |
my $max_bound = ( $last_n_minutes - 1 ); |
| 229 |
if ( $self->debug ) { |
| 230 |
warn "Check visits over the last $last_n_minutes minutes\n"; |
| 231 |
} |
| 232 |
foreach my $minus_minutes ( 0 .. $max_bound ) { |
| 233 |
my $timestamp = $now; |
| 234 |
if ($minus_minutes) { |
| 235 |
$timestamp -= ( $minus_minutes * 60 ); |
| 236 |
} |
| 237 |
my $expanded_time = scalar localtime($timestamp); |
| 238 |
if ( $self->debug ) { |
| 239 |
warn "Considered time slot: $expanded_time\n"; |
| 240 |
} |
| 241 |
my $cache_key = sprintf( "%s:%s", $ip_range, $timestamp ); |
| 242 |
push( @cache_keys, $cache_key ); |
| 243 |
} |
| 244 |
} |
| 245 |
} |
| 246 |
return \@cache_keys; |
| 247 |
} |
| 248 |
|
| 249 |
sub _prepare_template { |
| 250 |
my ( $self, $args ) = @_; |
| 251 |
my $template; |
| 252 |
my $interface = $self->interface; |
| 253 |
my $template_filename = $args->{template_filename}; |
| 254 |
if ( $interface && $template_filename ) { |
| 255 |
|
| 256 |
#NOTE: See Koha::Template in BZ 31380 |
| 257 |
$template = C4::Templates::gettemplate( $template_filename, $interface ); |
| 258 |
} |
| 259 |
return $template; |
| 260 |
} |
| 261 |
|
| 262 |
1; |