Use Capped collection on MongoAppender for Log4Net

One of the coolest feature of Mongo is the concept of Capped Collection, or “fixed size” collection. They are based on a FIFO queue where the first record to be discharded is the first inserted, and this is exceptional to create a log-collection that automatically purge all old logs without any user intervention.

To be able to automatically enable this feature on the Log4Net Mongo appender you need to do a little modification to the code, this is because the original code simply gets a reference to the collection with this code.

connection = MongoServer.Create(mongoConnectionString.ToString());
var db = connection.GetDatabase(DatabaseName);
collection = db.GetCollection(CollectionName);

C# drivers for Mongo automatically creates a collection if it is not present, this means that when you call db.GetCollection if the collection is not present it will be automatically created, but it is not capped. To solve this problem you can modify the initialization code with this code.

if (!db.CollectionExists(collectionName)) { 
    var options =  CollectionOptions
    db.CreateCollection(collectionName, options);
collection = db.GetCollection(CollectionName);

MongoDb C# drivers has a class called CollectionOptions used to setup options to create a new MongoCollection and it can be accessed with a really easy Fluent-Interface, in my example I call SetCapped(true) to enable a capped collection and SetMaxSize() to setup the maximum size in bytes. The size of the capped-collection is stored in the appender property called CappedSize, the default is 500MB, but you can setup any size you likes in standard log4Net configuration.

Gian Maria.

Published by

Ricci Gian Maria

.Net programmer, User group and community enthusiast, programmer - aspiring architect - and guitar player :). Visual Studio ALM MVP

2 thoughts on “Use Capped collection on MongoAppender for Log4Net”

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.