Is it possible in Cassandra to set the custom ttl values for specific information events to have these types of event deleted from the dbase after say a day or two? We have a very specific problem in that our system is designed around a large amount of services and one of the parameters of one of the service’s elements are changed very often and this causes an excessive amount of service update by scripting information events, some times 10s of thousands a day, that we don’t care about to be generated and fill up the info database table. This is causing a number of problems in MySQL such as the table becoming to large for DM to handle and also starving out all the very important user initiated command (what I refer to as system audit events) info events. With the default DM setting of limiting the info table to 300K events we often have the table only contain 3-5 days worth of events and %99 percent of these are useless unwanted events and in turn, we only have 3-5 days of records of what actions users have performed. We could increase the number of event the table holds but then we run into problems with the table being too big and starting to cause basic DM problems. I expect we will have the same or similar problem once we migrate to Cassandra. The best solution would be if DM had a setting to disable the logging of these types of events, but since that does not appear to be possible or something Skyline is willing to add the next best thing in Cassandra would be to set a short TTL for these types of events so the don’t fill up and load down the database. This single behavior is currently one of our biggest issues and is starting become a show stopper to delivering larger systems to our customers.

Any ideas? Thanks in advance