add server status check in cron_zkb startup

[lockefox]

http://eveonline-third-party-documentation.readthedocs.org/en/latest/xmlapi/serv_serversstatus/

If the EVE server isn't online, unlikely to be receiving new kills from zkb. Add test in "no kills received" logic chain?

Use eveapi (currently submodule in pcu_cron branch) to retrieve EVE XML API information

[lockefox]

NOTE:

• If CREST server is having issues, ZKB endpoint will not work
  • Use test kill? https://public-crest.eveonline.com/killmails/30290604/787fb3714062f1700560d4a83ce32c67640b1797/
• If EVE downtime, set retry loop to 0
  • If /eve/serverStatus:onlinePlayers < 100
  • if /eve/serverStatus:serverOpen == false

Still want to ask redisq endpoint if there are kills in this case, and send error email if no kills are written. Just simplify retry logic to be more robust against extended zkb downtimes