redis_logger¶ ↑

Application logging to redis.

Version 0.1.0

sudo gem install redis_logger

You will also want to install the web interface for viewing the log entries:

http://github.com/masonoise/redis_logger-web

1. Install redis (code.google.com/p/redis/) and redis_logger

2. In your application code, place a call to RedisLogger, using one of: debug, warn, error. Each call is expected to pass a Hash of data, and an optional list of groups to which you want the log entry to belong. For example, you might want to log all of your requests to a group called “requests”, so you might put this call into your application_controller.rb file in a before_filter:

RedisLogger.debug({ "request" => request.url,
                       "remote_ip" => request.remote_ip,
                       "user_id" => current_user.id,
                       "username" => current_user.email
}, "requests")

That would store the Hash containing four key/value pairs into the log entry, place it automatically in the “debug” group of entries, and also in the “requests” group of entries. Elsewhere in your code, you might log all user signup activity into a “signup” group, for example. Using the web interface, you can view the entries in a group or all the entries in an intersection of multiple groups – for example, all errors in the signup process.

If you want to use redis-cli to look things over instead of the web interface, here are the basics:

./redis-cli
sort logger:set:debug limit 0 10 desc
==> returns most recent 10 keys from the debug group, each of which maps to a Hash of log fields
hgetall log:1271274209
==> returns the keys and values for that log entry
smembers logger:sets
==> returns list of the groups that have been logged into

1. Currently the log keys are simply the longint timestamps of each entry. If multiple threads are going it’s very possible that they’ll log at the same instant and only one entry will survive. Need to work out a way to identify each server with a unique id that could be added to the log keys, such as log:1:<timestamp>, log:2:<timestamp>, etc.

2. It would be nice to cache the group names so it doesn’t have to add the group to logger:sets each time.

3. Support for viewing intersection of groups is limited to most recent 100 entries right now. It will be nice to keep the stored intersection set around for viewing, pagination, and export, but we need to have a way to clean up the stored sets periodically.

This is very young and simple thus far. If you have some cool ideas, go for it, give it a try, if it seems good to you then get in touch and we’ll see about merging it in.

• Fork the project.

• Make your feature addition or bug fix.

• Add tests for it. This is important so I don’t break it in a future version unintentionally.

• Commit, do not mess with rakefile, version, or history. (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)

• Send me a pull request. Bonus points for topic branches.

Mason Jones ~~~ stuff@charnel.com ~~~ @masonoise