If you are looking at building an Analytics engine/server look no further.
Querying these two sources was simple enough. If I wanted stateful information, I would query MySQL. If I wanted to know about user behaviour over time, query Splunk.
But when you needed to ask a question that involves both behaviour and state, such as how many users liked a photo in the last month (Splunk), broken down by user subscription type (MySQL), you would need an offline way to query the two data sources separately and then combine the two. For this, I used Python with its MySQL and Splunk libraries.
This was a minor headache.