I've been told missing blocks often has to do with where a BP is scheduled in the list, and less to do with an individual BPs infrastructure. For quite some time eosdacserver was scheduled right after eoscleanerbp which had some very high latency. The scheduler is done alphabetically right now, so eosdacserver comes after eoscleanerbp which can cause mini forks and missed blocks. See an example of log output here.
You can also see the schedule is alphabetical here:
promoting proposed schedule (set in block 6888584) to pending; current block: 6888904 lib: 6888588 schedule:
{"version":173,"producers":[
{"producer_name":"argentinaeos","block_signing_key":"EOS7jq4FHrFrtCXxpRQ39dBeDMa5AjM4VaRbqBECkSa5aZnizJzrx"},
{"producer_name":"bitfinexeos1","block_signing_key":"EOS6sgKjHUFtY1XxxQaMDwfxBac6nDBibVzZHb8LFMVmvSjcCdDhE"},
{"producer_name":"eos42freedom","block_signing_key":"EOS4tw7vH62TcVtMgm2tjXzn9hTuHEBbGPUK2eos42ssY7ip4LTzu"},
{"producer_name":"eosamsterdam","block_signing_key":"EOS8HDwgQnKMHAxzva9eyHsPeETv23JSQ6RRzQjJLvCsZ6b1ySgQo"},
{"producer_name":"eosauthority","block_signing_key":"EOS4va3CTmAcAAXsT26T3EBWqYHgQLshyxsozYRgxWm9tjmy17pVV"},
{"producer_name":"eosbixinboot","block_signing_key":"EOS7QC1XfAtkYeLjbHQjcDWVqUsxuSJ3YRhNyG93VAv2u3uHopGVp"},
{"producer_name":"eoscafeblock","block_signing_key":"EOS7MAPWVuYcxNtc2n9e6WaEedEZd9thGVHn2Wpu2PoMhNiteTTqL"},
{"producer_name":"eoscanadacom","block_signing_key":"EOS5HYV7rWeRxpZMCooe8YHRFQHKK7ncdmmUMTe3wCMaY2EvyVzUx"},
{"producer_name":"eosdacserver","block_signing_key":"EOS6VkageCqyz1X97n7E1JyjPoZu8HEX6CVujY5ABzk3fPyV3vzZK"},
{"producer_name":"eoshuobipool","block_signing_key":"EOS5XKswW26cR5VQeDGwgNb5aixv1AMcKkdDNrC59KzNSBfnH6TR7"},
{"producer_name":"eosisgravity","block_signing_key":"EOS55HTTjoxVX1zVpW8pabxygBb1J3SEnG5D8D3y3KgrnSbLpELfE"},
{"producer_name":"eoslaomaocom","block_signing_key":"EOS8QgURqo875qu3a8vgZ58qBeu2cTehe9zAWRfpdCXAQipicu1Fi"},
{"producer_name":"eosliquideos","block_signing_key":"EOS4v1n2j5kXbCum8LLEc8zQLpeLK9rKVFmsUgLCWgMDN38P6PcrW"},
{"producer_name":"eosnewyorkio","block_signing_key":"EOS6GVX8eUqC1gN1293B3ivCNbifbr1BT6gzTFaQBXzWH9QNKVM4X"},
{"producer_name":"eosriobrazil","block_signing_key":"EOS7RioGoHQnhv2fJEiciP9Q7J8JgfJYFcyofVkmCqMop8Q1PzgqP"},
{"producer_name":"eosswedenorg","block_signing_key":"EOS7SGSBsWhSob6TEric6u3TGodcc1uXFcqSrquJ3Etuqcbb3VnNY"},
{"producer_name":"helloeoscnbp","block_signing_key":"EOS79cHpaEittzgJWgj79tdRhgzLEWy8wXmmQ3fL7kkDjmYYiGNet"},
{"producer_name":"jedaaaaaaaaa","block_signing_key":"EOS5XP49jx9eJNQZjqvP7pfKNaAe5JmyAbPZMdsiqD61QQEExqUYe"},
{"producer_name":"libertyblock","block_signing_key":"EOS7TqVs7LpMzJwikiWkGTG9UGSwcQVhxP2ZKBM7FCbUwcrryNgHP"},
{"producer_name":"starteosiobp","block_signing_key":"EOS5mB5dYCpQTHUPBS8SJ52sidTY9t7QNsZ2wg1Nzo5a1jKQM3Qx4"},
{"producer_name":"zbeosbp11111","block_signing_key":"EOS7rhgVPWWyfMqjSbNdndtCK8Gkza3xnDbUupsPLMZ6gjfQ4nX81"}
]}
I asked Dan about this:
And he seemed to think it's normal:
So maybe we should be talking more about latency and less about missed blocks? I did a little analysis of latency as well you may find interesting.
You are probably right on the latency issue , it was something I was concerned some time ago , but did not look to be an issue till now
This report dont really show missed blocks , but produced blocks , I know it is a thick approach , but may alert on some problems
I am working on make a more accurate reports , based on the blocks a BP should have produced for each production schedule change .I am setting up the mongo database and will code the logic