How accurate is awstats?

Very different results.

[imagerage]

How accurate is the tracking for awstats?

The difference in hits between awstats and webalizer is 10 times as much? Let's say there are 10000 hits in awstats, webalizer would show 1000 hits. I don't think they should be so far apart or are both of them not accurate? Do I need a paid solution to track correctly?

[gregbo]

Can you create some test log files with a known number of hits, pageviews, etc. and feed them to both programs to get an apples to apples comparison?

[imagerage]

How do I do that? Sorry I am a real newbie.

[gregbo]

Do you have access to the log files? Also, are you using something Unix-like? The basic idea is to take a snippet of log data (or generate a text file containing a small number of simple GET requests) and feed it to both programs. Keep adding additional data to the log files until you notice a difference between the outputs of both programs.

One thing I noticed is that sometimes awstats won't report on log data with timestamps that are out of order.

[imagerage]

I think I can get the raw logs.

Have your awstats ever skip stats?

[gregbo]

Have your awstats ever skip stats?

Actually, I was just using it on some test data to see if it would work for another purpose. The summary report indicated corrupted data. I checked the source code and discovered that under some conditions, awstats will report corrupted data when timestamps are out of order.

If your logs come from multiple servers, some of which may be delayed (e.g. due to nonrotating logs), you might have out of order data.