实时绘制应用程序事件指标的解决方案
我们有一个解析推文的应用程序，我们希望实时查看活动。我们尝试了几种解决方案但没有成功。我们的主要问题是图形解决方案（例如：石墨）需要连续的指标流。当数据库聚合指标时，它是完成的平均操作，而不是求和。
我们最近看到了 cube from square ，它符合我们的要求，但它太新了。
还有其他选择吗？
We have an application that parses tweets and we want to see the activity in real time. We have tried several solution without success. Our main problems is that the graphing solution (example:graphite), needs a continious flow of metrics. When the db aggregates the metrics it's an average operation which is done, not a a sum.
We recently saw cube from square which would fit our requirement but it's too new.
Any alternatives?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
我在石墨的最新版本中找到了解决方案：
http://graphite.readthedocs.org/en/latest /config-carbon.html#storage-aggregation-conf
I found the solution in the last version of graphite:
http://graphite.readthedocs.org/en/latest/config-carbon.html#storage-aggregation-conf
如果我理解正确的话，你不能实时提供石墨，例如，一旦你发现一条新推文？
如果是这样的话，看起来你可以在更新石墨时指定一个unix时间戳
metric_path 值时间戳\n这样您就可以传入发现/发布/任何内容的时间，无论您何时处理它。If I understood correctly, you cannot feed graphite in realtime, for instance as soon as you discover a new tweet?
If that's the case, it looks like you can specify a unix timestamp when updating graphite
metric_path value timestamp\nso you could pass in the time of discovery/publication/whatever, regardless of when you process it.