system_health 0.0.5 → 0.0.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 649b6b51406bda1aec88b5865610daaf7886daf1
-  data.tar.gz: 1155745115676abe0b455e8d672d97477594f958
+  metadata.gz: 64fc6a4ae13390505cc788003d0689e4852b9a65
+  data.tar.gz: 079ac177a733ba941db774c4c4b5232c48358b8f
 SHA512:
-  metadata.gz: ce6bbd99f24675ef3e6b6a522905ee43cd73c5f406666cd9eefadf9cd97752a960ae5ade7f6177efe35517427c8ee72aeccf7eb5d65ca0719d5a1be01ac77359
-  data.tar.gz: 139474d30143f8904c40ad0fc6e3ad86aace8e76a01878fafdc05bf3e27fc67eb310829021085131a4037bb3b158b969490d1fd784f278bc1b8340b41178c9c9
+  metadata.gz: fa659c4506bf06cb2dede60f15166187802e239ffbe407282ca1e38165dcb89a92c759efd8e6b953f25805dbff4552d208d3e6328453a0b6be02986c177aed5e
+  data.tar.gz: db1c220f31218da8bb7f7d5fd0365dbf5edd0a9bb6727292ab0e55915197a59fc19335b64dc40ec6cec3e0f7491108884916a517f360594f3b0d7c041900e604

data/README.md

@@ -2,11 +2,7 @@
 
 The System Health gem can be added to your Rails application to provide
 a convenient way to regularly look for bad data or other system health
-indicators.  It provides a single endpoint that will generate an error
-count and enumerate error messages.  This count and the messages can be
-collected by external monitoring tools that look for the HTTP status code
-(200 when there are no errors or 500 when there is at least one error)
-and/or inspect the JSON payload that is returned.
+indicators.
 
 ## Installation
 
@@ -24,81 +20,98 @@ Or install it yourself as:
 
 and then...
 
-1. create an initializer in config/initializers/system_health.rb
+1. Create your monitor classes
+
+  All monitoring classes require a public instance method named `description`.
+  This method should return a string describing the bad data condition that is
+  being tested.
+
+## Generic monitoring class
+
+  For example, in lib/system_health/monitors/bad_data.rb create:
 
   ```ruby
-  SystemHealth.configure do |config|
-    config.monitor_classes = [
-      SystemHealth::Monitors::BadData
-    ]
+  module SystemHealth
+    module Monitors
+      class BadData < Base
+        def description
+          'Bad data was discovered'
+        end
+
+        private
+
+        def bad_data?
+          # return true from this method if there is bad data
+        end
+      end
+    end
   end
   ```
 
-  This should define the SYSTEM_HEALTH_MONITOR_CLASSES array with all
-  your monitor classes.  There can be as many monitor classes as you wish.
+  define a private instance method `bad_data?` that should return true
+  when bad data exists.
 
-2. Create your monitor classes
+## SQL monitoring class
 
-  In lib/system_health/monitors/bad_data.rb create:
+  For example, in lib/system_health/monitors/bad_sql_data.rb create:
 
   ```ruby
   module SystemHealth
     module Monitors
-      class BadData < Base
-        def error_messages
-          ['some error message here'] if bad_data?
+      class BadSqlData < Sql
+        def description
+          'Bad data was discovered through a SQL query'
         end
 
         private
 
-        def bad_data?
-          # logic here
+        def sql
+          <<-SQL
+            SELECT *
+            FROM some_table
+            WHERE bad_data is true
+          SQL
         end
       end
     end
   end
   ```
 
-  Each monitor should define a public instance method for `error_messages`
-  as an array.  If that array is empty then the data is good.  When the
-  data is bad, one or more error messages can be placed in that array.
-  That is it!
+  this type of monitor requires an additional private instance method
+  named `sql`.  This SQL statement should return rows for any data
+  integrity problem.  I.e. no rows means no problem.  Rows returned means
+  there is a problem.
 
-3. Define the environment variables
 
-  For applications hosted on Heroku, this might be something like:
+2. create an initializer in config/initializers/system_health.rb
 
-  ```bash
-  heroku config:add SYSTEM_HEALTH_MONITOR_USERNAME=somenamehere
-  heroku config:add SYSTEM_HEALTH_MONITOR_PASSWORD=somesecrethere
+  ```ruby
+  SystemHealth.configure do |config|
+    config.monitor_classes = [
+      SystemHealth::Monitors::BadData,
+      SystemHealth::Monitors::BadSqlData
+    ]
+  end
   ```
 
-  on other environments, you'll need to define these elsewhere.
-
 ## Usage
 
-The System Health gem exposes a single controller endpoint at:
-/system_health/monitor
-
-It will return a 200 HTTP status code if there are no errors and a 500
-HTTP status code if there are any errors.  It also returns a JSON
-payload with data like this:
+The System Health gem exposes a monitoring class `SystemHealth::Monitor` that
+can be used to test for all system health issues with the following
+methods:
 
-```json
-{
-  "error_count": 2,
-  "messages": ["some error message here", "another error message"]
-}
+```ruby
+mon = SystemHealth::Monitor.new
+mon.error_messages
+mon.error_count
 ```
 
 ## To do?
 
-1. Consider how to deal with long running monitors that may fail with
-   page load limits (e.g. Heroku's 30 second timeout)
-2. Add ability to run different health monitors at different times or
-   possibly call monitors individually
-3. Add the ability to generate error notifications directly from System
-   Health instead of relying on external monitoring tools to send these
+1. Add concept of notifiers to make it more seamless to email or report
+   on system health issues.
+2. Add monitor classes so different classes can be run at different
+   times and with different frequency.
 
 ## Contributing
 

data/lib/system_health/monitor.rb

@@ -1,6 +1,6 @@
 module SystemHealth
   class Monitor
-    def initialize(monitor_classes)
+    def initialize(monitor_classes = SystemHealth.configuration.monitor_classes)
       @monitor_classes = monitor_classes
     end
 

data/lib/system_health/monitors/sql.rb

@@ -0,0 +1,17 @@
+module SystemHealth
+  module Monitors
+    class Sql < Base
+
+      private
+
+      def bad_data?
+        ActiveRecord::Base.connection.execute(sql).count > 0
+      end
+
+      def sql
+        raise
+      end
+    end
+  end
+end
+

data/lib/system_health/version.rb

@@ -1,3 +1,3 @@
 module SystemHealth
-  VERSION = "0.0.5"
+  VERSION = "0.0.7"
 end

data/spec/lib/system_health/monitor_spec.rb

@@ -7,8 +7,14 @@ module SystemHealth
 
     context 'when errors' do
       class MonitorDouble < Monitors::Base
-        def add_error_messages
-          add_error_message('hello')
+        def description
+          'hello'
+        end
+
+        private
+
+        def bad_data?
+          true
         end
       end
 
@@ -36,7 +42,14 @@ module SystemHealth
 
     context 'when no errors' do
       class MonitorDouble < Monitors::Base
-        def add_error_messages
+        def description
+          'hello'
+        end
+
+        private
+
+        def bad_data?
+          false
         end
       end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: system_health
 version: !ruby/object:Gem::Version
-  version: 0.0.5
+  version: 0.0.7
 platform: ruby
 authors:
 - Stirling Olson
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-12-28 00:00:00.000000000 Z
+date: 2016-12-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rails
@@ -85,6 +85,7 @@ files:
 - lib/system_health/engine.rb
 - lib/system_health/monitor.rb
 - lib/system_health/monitors/base.rb
+- lib/system_health/monitors/sql.rb
 - lib/system_health/version.rb
 - spec/lib/system_health/monitor_spec.rb
 - system_health.gemspec