Running multiple instances about tyk

[lonelycode]

Hi Alex,

I agree that option 3 sounds the most attractive, as it also opens up the data collection element to third-party processes in case Mongo isn't suitable back-end (Mongo is used mainly because of the aggregation framework and the Tyk Dashboard).

Personally I think simpler is better :-)

It should be relatively easy to enable this by simply checking the purgeDelay value in the globals setup and not firing off the goroutine (with an appropriate log message, we'll need to warn the user that their Redis DB might get massive), though I might use -1 instead of 0?

https://github.com/lonelycode/tyk/blob/86678f56ee7ab27194e8e53aad3c9a7fa8422a78/main.go#L55

Longer term perhaps the analytics handler should be converted to an interface, that way Tyk could write directly to alternative stores...

Will add that to roadmap.

Cheers,
Martin

[alexwhitman]

Thanks, I've implemented disabling of the cache purge in #9.

[lonelycode]

Fixed in #9