Hi, I am using memcached 1.4.4 with Django 1.1, FastCGI and Lighttpd. I have the django-memcached_status app installed so I can monitor stats. I am using three memcached daemons per site, each on a different physical server (one on localhost with minimal allocated RAM, the others on remote servers).
My issue is that I don't know how to interprete the stats to see if I need to tune my memcached daemons or not. I'm getting between 3% - 15% misses on daemons, but even at busy times the load % shown in django admin shows as 0% - is this a bug in django- memcached_status or is there something wrong with my memcached install? Also, I can see that the bytes written value keeps climbing (it's at well over 30GB on some daemons which have been running for a few days), yet the bytes read value is always much much smaller (curently less than 1GB even on the busiest site). Is this indicative of a problem with how caching has been implemented in our code, ie: if the cache is being rewritten very frequently, as seems the case? My cache expiry time is set to 3 mins. Any advice greatly received! I feel like I've loads of info, but no clue as to what to do with it. Simon.
