connection pooling - Connecting to Azure Cache Service takes about 3.3 seconds -
in our still-in-development project have noticed sudden delays when accessing our asp.net web api services. using awesome mini profiler nailed these delays caused when connections azure data cache (preview) services dropped , have reestablished. process takes 3.3 seconds. after reconnecting, getting object cache takes 1.4 ms.
when increased maxconnectionstoserver 1 20, noticed thing. if don't make requests web api 1 or 2 minutes (that's when connections dropped) , start making calls, next 20 requests delayed 3.3 seconds, how connection pooling works guess (round-tripping connections pool).
both web api , caching service hosted in east region, have disabled local cache, ssl disabled, auto discover enabled.
so, i'm wondering if wrong our configuration or thing because azure cache still in preview?
any information valued.
thanks!
it sounds shared cache being offloaded due inactivity. 1 way test add in-role cache existing service (if available) , swap cache usage new cache. in-role cache described here.
once cache moved off of shared offering, wait requisite 1-2 minutes idle time out , retry connection, delay should not present.
assuming want stick shared cache option after isolating problem, current workaround aware of running background task periodically ping cache keep alive.
if running full web role can launch background task on application start up.
if deploying via mobile services, can run "ping" via scheduled jobs. issue may run here minimum time scheduled job 1 minute, may not aggressive enough keep cache alive 100% of time.
Comments
Post a Comment