Add rate limiting and input validation security enhancements

lib/vmpooler/api/helpers.rb

@@ -1,10 +1,13 @@
 # frozen_string_literal: true
 
+require 'vmpooler/api/input_validator'
+
 module Vmpooler
 
   class API
 
     module Helpers
+      include InputValidator
 
       def tracer
         @tracer ||= OpenTelemetry.tracer_provider.tracer('api', Vmpooler::VERSION)

lib/vmpooler/api/input_validator.rb

@@ -0,0 +1,159 @@
+# frozen_string_literal: true
+
+module Vmpooler
+  class API
+    # Input validation helpers to enhance security
+    module InputValidator
+      # Maximum lengths to prevent abuse
+      MAX_HOSTNAME_LENGTH = 253
+      MAX_TAG_KEY_LENGTH = 50
+      MAX_TAG_VALUE_LENGTH = 255
+      MAX_REASON_LENGTH = 500
+      MAX_POOL_NAME_LENGTH = 100
+      MAX_TOKEN_LENGTH = 64
+
+      # Valid patterns
+      HOSTNAME_PATTERN = /\A[a-z0-9]([a-z0-9-]{0,61}[a-z0-9])?(\.[a-z0-9]([a-z0-9-]{0,61}[a-z0-9])?)* \z/ix.freeze
+      POOL_NAME_PATTERN = /\A[a-zA-Z0-9_-]+\z/.freeze
+      TAG_KEY_PATTERN = /\A[a-zA-Z0-9_\-.]+\z/.freeze
+      TOKEN_PATTERN = /\A[a-zA-Z0-9\-_]+\z/.freeze
+      INTEGER_PATTERN = /\A\d+\z/.freeze
+
+      class ValidationError < StandardError; end
+
+      # Validate hostname format and length
+      def validate_hostname(hostname)
+        return error_response('Hostname is required') if hostname.nil? || hostname.empty?
+        return error_response('Hostname too long') if hostname.length > MAX_HOSTNAME_LENGTH
+        return error_response('Invalid hostname format') unless hostname.match?(HOSTNAME_PATTERN)
+
+        true
+      end
+
+      # Validate pool/template name
+      def validate_pool_name(pool_name)
+        return error_response('Pool name is required') if pool_name.nil? || pool_name.empty?
+        return error_response('Pool name too long') if pool_name.length > MAX_POOL_NAME_LENGTH
+        return error_response('Invalid pool name format') unless pool_name.match?(POOL_NAME_PATTERN)
+
+        true
+      end
+
+      # Validate tag key and value
+      def validate_tag(key, value)
+        return error_response('Tag key is required') if key.nil? || key.empty?
+        return error_response('Tag key too long') if key.length > MAX_TAG_KEY_LENGTH
+        return error_response('Invalid tag key format') unless key.match?(TAG_KEY_PATTERN)
+
+        if value
+          return error_response('Tag value too long') if value.length > MAX_TAG_VALUE_LENGTH
+
+          # Sanitize value to prevent injection attacks
+          sanitized_value = value.gsub(/[^\w\s\-.@:\/]/, '')
+          return error_response('Tag value contains invalid characters') if sanitized_value != value
+        end
+
+        true
+      end
+
+      # Validate token format
+      def validate_token_format(token)
+        return error_response('Token is required') if token.nil? || token.empty?
+        return error_response('Token too long') if token.length > MAX_TOKEN_LENGTH
+        return error_response('Invalid token format') unless token.match?(TOKEN_PATTERN)
+
+        true
+      end
+
+      # Validate integer parameter
+      def validate_integer(value, name = 'value', min: nil, max: nil)
+        return error_response("#{name} is required") if value.nil?
+
+        value_str = value.to_s
+        return error_response("#{name} must be a valid integer") unless value_str.match?(INTEGER_PATTERN)
+
+        int_value = value.to_i
+        return error_response("#{name} must be at least #{min}") if min && int_value < min
+        return error_response("#{name} must be at most #{max}") if max && int_value > max
+
+        int_value
+      end
+
+      # Validate VM request count
+      def validate_vm_count(count)
+        validated = validate_integer(count, 'VM count', min: 1, max: 100)
+        return validated if validated.is_a?(Hash) # error response
+
+        validated
+      end
+
+      # Validate disk size
+      def validate_disk_size(size)
+        validated = validate_integer(size, 'Disk size', min: 1, max: 2048)
+        return validated if validated.is_a?(Hash) # error response
+
+        validated
+      end
+
+      # Validate lifetime (TTL) in hours
+      def validate_lifetime(lifetime)
+        validated = validate_integer(lifetime, 'Lifetime', min: 1, max: 168) # max 1 week
+        return validated if validated.is_a?(Hash) # error response
+
+        validated
+      end
+
+      # Validate reason text
+      def validate_reason(reason)
+        return true if reason.nil? || reason.empty?
+        return error_response('Reason too long') if reason.length > MAX_REASON_LENGTH
+
+        # Sanitize to prevent XSS/injection
+        sanitized = reason.gsub(/[<>"']/, '')
+        return error_response('Reason contains invalid characters') if sanitized != reason
+
+        true
+      end
+
+      # Sanitize JSON body to prevent injection
+      def sanitize_json_body(body)
+        return {} if body.nil? || body.empty?
+
+        begin
+          parsed = JSON.parse(body)
+          return error_response('Request body must be a JSON object') unless parsed.is_a?(Hash)
+
+          # Limit depth and size to prevent DoS
+          return error_response('Request body too complex') if json_depth(parsed) > 5
+          return error_response('Request body too large') if body.length > 10_240 # 10KB max
+
+          parsed
+        rescue JSON::ParserError => e
+          error_response("Invalid JSON: #{e.message}")
+        end
+      end
+
+      # Check if validation result is an error
+      def validation_error?(result)
+        result.is_a?(Hash) && result['ok'] == false
+      end
+
+      private
+
+      def error_response(message)
+        { 'ok' => false, 'error' => message }
+      end
+
+      def json_depth(obj, depth = 0)
+        return depth unless obj.is_a?(Hash) || obj.is_a?(Array)
+        return depth + 1 if obj.empty?
+
+        if obj.is_a?(Hash)
+          depth + 1 + obj.values.map { |v| json_depth(v, 0) }.max
+        else
+          depth + 1 + obj.map { |v| json_depth(v, 0) }.max
+        end
+      end
+    end
+  end
+end

lib/vmpooler/api/rate_limiter.rb

@@ -0,0 +1,116 @@
+# frozen_string_literal: true
+
+module Vmpooler
+  class API
+    # Rate limiter middleware to protect against abuse
+    # Uses Redis to track request counts per IP and token
+    class RateLimiter
+      DEFAULT_LIMITS = {
+        global_per_ip: { limit: 100, period: 60 }, # 100 requests per minute per IP
+        authenticated: { limit: 500, period: 60 }, # 500 requests per minute with token
+        vm_creation: { limit: 20, period: 60 },    # 20 VM creations per minute
+        vm_deletion: { limit: 50, period: 60 }     # 50 VM deletions per minute
+      }.freeze
+
+      def initialize(app, redis, config = {})
+        @app = app
+        @redis = redis
+        @config = DEFAULT_LIMITS.merge(config[:rate_limits] || {})
+        @enabled = config.fetch(:rate_limiting_enabled, true)
+      end
+
+      def call(env)
+        return @app.call(env) unless @enabled
+
+        request = Rack::Request.new(env)
+        client_id = identify_client(request)
+        endpoint_type = classify_endpoint(request)
+
+        # Check rate limits
+        return rate_limit_response(client_id, endpoint_type) if rate_limit_exceeded?(client_id, endpoint_type, request)
+
+        # Track the request
+        increment_request_count(client_id, endpoint_type)
+
+        @app.call(env)
+      end
+
+      private
+
+      def identify_client(request)
+        # Prioritize token-based identification for authenticated requests
+        token = request.env['HTTP_X_AUTH_TOKEN']
+        return "token:#{token}" if token && !token.empty?
+
+        # Fall back to IP address
+        ip = request.ip || request.env['REMOTE_ADDR'] || 'unknown'
+        "ip:#{ip}"
+      end
+
+      def classify_endpoint(request)
+        path = request.path
+        method = request.request_method
+
+        return :vm_creation if method == 'POST' && path.include?('/vm')
+        return :vm_deletion if method == 'DELETE' && path.include?('/vm')
+        return :authenticated if request.env['HTTP_X_AUTH_TOKEN']
+
+        :global_per_ip
+      end
+
+      def rate_limit_exceeded?(client_id, endpoint_type, _request)
+        limit_config = @config[endpoint_type] || @config[:global_per_ip]
+        key = "vmpooler__ratelimit__#{endpoint_type}__#{client_id}"
+
+        current_count = @redis.get(key).to_i
+        current_count >= limit_config[:limit]
+      rescue StandardError => e
+        # If Redis fails, allow the request through (fail open)
+        warn "Rate limiter Redis error: #{e.message}"
+        false
+      end
+
+      def increment_request_count(client_id, endpoint_type)
+        limit_config = @config[endpoint_type] || @config[:global_per_ip]
+        key = "vmpooler__ratelimit__#{endpoint_type}__#{client_id}"
+
+        @redis.pipelined do |pipeline|
+          pipeline.incr(key)
+          pipeline.expire(key, limit_config[:period])
+        end
+      rescue StandardError => e
+        # Log error but don't fail the request
+        warn "Rate limiter increment error: #{e.message}"
+      end
+
+      def rate_limit_response(client_id, endpoint_type)
+        limit_config = @config[endpoint_type] || @config[:global_per_ip]
+        key = "vmpooler__ratelimit__#{endpoint_type}__#{client_id}"
+
+        begin
+          ttl = @redis.ttl(key)
+        rescue StandardError
+          ttl = limit_config[:period]
+        end
+
+        headers = {
+          'Content-Type' => 'application/json',
+          'X-RateLimit-Limit' => limit_config[:limit].to_s,
+          'X-RateLimit-Remaining' => '0',
+          'X-RateLimit-Reset' => (Time.now.to_i + ttl).to_s,
+          'Retry-After' => ttl.to_s
+        }
+
+        body = JSON.pretty_generate({
+                                      'ok' => false,
+                                      'error' => 'Rate limit exceeded',
+                                      'limit' => limit_config[:limit],
+                                      'period' => limit_config[:period],
+                                      'retry_after' => ttl
+                                    })
+
+        [429, headers, [body]]
+      end
+    end
+  end
+end

lib/vmpooler/api/v3.rb

@@ -1085,9 +1085,29 @@ module Vmpooler
         result = { 'ok' => false }
         metrics.increment('http_requests_vm_total.post.vm.checkout')
 
-        payload = JSON.parse(request.body.read)
+        # Validate and sanitize JSON body
+        payload = sanitize_json_body(request.body.read)
+        if validation_error?(payload)
+          status 400
+          return JSON.pretty_generate(payload)
+        end
 
-        if payload
+        # Validate each template and count
+        payload.each do |template, count|
+          validation = validate_pool_name(template)
+          if validation_error?(validation)
+            status 400
+            return JSON.pretty_generate(validation)
+          end
+
+          validated_count = validate_vm_count(count)
+          if validation_error?(validated_count)
+            status 400
+            return JSON.pretty_generate(validated_count)
+          end
+        end
+
+        if payload && !payload.empty?
           invalid = invalid_templates(payload)
           if invalid.empty?
             result = atomically_allocate_vms(payload)
@@ -1206,6 +1226,7 @@ module Vmpooler
         result = { 'ok' => false }
         metrics.increment('http_requests_vm_total.get.vm.template')
 
+        # Template can contain multiple pools separated by +, so validate after parsing
         payload = extract_templates_from_query_params(params[:template])
 
         if payload
@@ -1235,6 +1256,13 @@ module Vmpooler
         status 404
         result['ok'] = false
 
+        # Validate hostname
+        validation = validate_hostname(params[:hostname])
+        if validation_error?(validation)
+          status 400
+          return JSON.pretty_generate(validation)
+        end
+
         params[:hostname] = hostname_shorten(params[:hostname])
 
         rdata = backend.hgetall("vmpooler__vm__#{params[:hostname]}")
@@ -1373,6 +1401,13 @@ module Vmpooler
         status 404
         result['ok'] = false
 
+        # Validate hostname
+        validation = validate_hostname(params[:hostname])
+        if validation_error?(validation)
+          status 400
+          return JSON.pretty_generate(validation)
+        end
+
         params[:hostname] = hostname_shorten(params[:hostname])
 
         rdata = backend.hgetall("vmpooler__vm__#{params[:hostname]}")
@@ -1403,16 +1438,21 @@ module Vmpooler
 
         failure = []
 
+        # Validate hostname
+        validation = validate_hostname(params[:hostname])
+        if validation_error?(validation)
+          status 400
+          return JSON.pretty_generate(validation)
+        end
+
         params[:hostname] = hostname_shorten(params[:hostname])
 
         if backend.exists?("vmpooler__vm__#{params[:hostname]}")
-          begin
-            jdata = JSON.parse(request.body.read)
-          rescue StandardError => e
-            span = OpenTelemetry::Trace.current_span
-            span.record_exception(e)
-            span.status = OpenTelemetry::Trace::Status.error(e.to_s)
-            halt 400, JSON.pretty_generate(result)
+          # Validate and sanitize JSON body
+          jdata = sanitize_json_body(request.body.read)
+          if validation_error?(jdata)
+            status 400
+            return JSON.pretty_generate(jdata)
           end
 
           # Validate data payload
@@ -1421,6 +1461,13 @@ module Vmpooler
               when 'lifetime'
                 need_token! if Vmpooler::API.settings.config[:auth]
 
+                # Validate lifetime is a positive integer
+                lifetime_int = arg.to_i
+                if lifetime_int <= 0
+                  failure.push("Lifetime must be a positive integer (got #{arg})")
+                  next
+                end
+
                 # in hours, defaults to one week
                 max_lifetime_upper_limit = config['max_lifetime_upper_limit']
                 if max_lifetime_upper_limit
@@ -1430,13 +1477,17 @@ module Vmpooler
                   end
                 end
 
-                # validate lifetime is within boundaries
-                unless arg.to_i > 0
-                  failure.push("You provided a lifetime (#{arg}) but you must provide a positive number.")
-                end
-
               when 'tags'
                 failure.push("You provided tags (#{arg}) as something other than a hash.") unless arg.is_a?(Hash)
+
+                # Validate each tag key and value
+                arg.each do |key, value|
+                  tag_validation = validate_tag(key, value)
+                  if validation_error?(tag_validation)
+                    failure.push(tag_validation['error'])
+                  end
+                end
+
                 failure.push("You provided unsuppored tags (#{arg}).") if config['allowed_tags'] && !(arg.keys - config['allowed_tags']).empty?
               else
                 failure.push("Unknown argument #{arg}.")
@@ -1478,9 +1529,23 @@ module Vmpooler
         status 404
         result = { 'ok' => false }
 
+        # Validate hostname
+        validation = validate_hostname(params[:hostname])
+        if validation_error?(validation)
+          status 400
+          return JSON.pretty_generate(validation)
+        end
+
+        # Validate disk size
+        validated_size = validate_disk_size(params[:size])
+        if validation_error?(validated_size)
+          status 400
+          return JSON.pretty_generate(validated_size)
+        end
+
         params[:hostname] = hostname_shorten(params[:hostname])
 
-        if ((params[:size].to_i > 0 )and (backend.exists?("vmpooler__vm__#{params[:hostname]}")))
+        if backend.exists?("vmpooler__vm__#{params[:hostname]}")
           result[params[:hostname]] = {}
           result[params[:hostname]]['disk'] = "+#{params[:size]}gb"