investigate: plugins showing up as inactive upon redeploy

[edaniszewski]

this could be an intermittent issue due to a cluster with a bit of development churn, but something to keep an eye on. when a set of plugins were redeployed and a plugin refresh was manually kicked off (via /plugin?refresh=true), it looks like the plugins were still in the inactive state, though there did not appear to be any reason for this in both the synse server logs and in the plugin logs.

as there was nothing conclusive in the logs, it would be good to cycle plugins a bit to try and determine reproducibility conditions. if reproducible, perhaps adding more logs is warranted for greater visibility into what is going on

[edaniszewski]

I believe this could also have detrimental effects to device caching. If a plugin is inactive, it will not be scanned for its devices on cache rebuild, effectively ghosting the device from the system for a while (e.g. until the plugin is marked active again and the device cache is rebuilt). This would also explain why in the above scenario, various devices were missing until a server restart occurred.