Don't rely on Logs for Important Info
In another post you said that you shouldn't rely on logs for business critical information. Why is that? I am doing research on stats and I need them to be as accurate as possible.I was planning on using S3stat to track my bandwidth cost per file stored on S3. The bandwidth costs are then going to be passed off to others on a file-by-file basis. (I will have under 500 files stored that range from 20MB to 1GB.)Thank you for your help!
Tuesday, November 11, 2008
I think it's generally in your best interest to keep your mission critical processes under direct control. In the case of the previous poster, and possibly yours too, it might make more sense build a custom reporting tool that exactly suits your needs.If you absolutely need to account for every file transfer, I would recommend passing them through some form of server-side redirect that can log requests directly. You could then compare those request numbers with the raw logfiles provided by S3 to get a better feel for which downloads completed and which were abandoned.While you could certainly accomplish that second task simply by reading the reports generated by S3stat, there would likely be times where you would appreciate having finer control over the reporting process. At the end of the day, it's really not that difficult to duplicate the portion of our service that you need for this business case. And it's probably good advice in general not to bet your whole business on a 3rd party service that's charging you two bucks a month :)Regardless, I'd recommend signing up for a free trial if you've not yet done so. Best case, you might find it meets your needs perfectly. Worst case, you'll probably find that it makes for a good sanity check against your home-grown tools.Good luck!
Friday, November 14, 2008
[ reply to this topic ]
[ return to topic list ]