DEAR: Distributed Evaluation of Alerting Rules
DEAR: Distributed Evaluation of Alerting Rules
复制标题
DOI：
10.1109/cloud49709.2020.00034
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
M. Mormul;Pascal Hirmer;Christoph Stach;B. Mitschang
中科院分区：
文献类型：
--
作者：
M. Mormul;Pascal Hirmer;Christoph Stach;B. Mitschang
云计算很久以前就走过了炒作周期，从那时起就牢固地确立了自己作为未来技术的地位。然而，为了尽可能经济高效地利用云，持续监控是防止资源过度或不足调试的关键。在大规模场景中，云监控面临的一些挑战，如高网络流量、低监控数据准确性和高洞察时间，需要在考虑管理复杂性的同时，在IT运营中采用新的方法。为了应对这些挑战，我们提出了警报规则的分布式评估（DEAR）。DEAR是一个用于监控系统的插件，它自动将警报规则分配给被监控的资源，以解决高精度和低网络流量之间的权衡，而无需管理开销。我们根据当今IT监控的需求评估我们的方法，并将其与传统的基于代理的监控方法进行比较。
Cloud computing passed the hype cycle long ago and firmly established itself as a future technology since then. However, to utilize the cloud as cost-efficiently as possible, a continuous monitoring is key to prevent an over- or under-commissioning of resources. In large-scaled scenarios, several challenges for cloud monitoring, such as high network traffic volume, low accuracy of monitoring data, and high time-to-insight, require new approaches in IT Operations while considering administrative complexity. To handle these challenges, we present DEAR, the Distributed Evaluation of Alerting Rules. DEAR is a plugin for monitoring systems which automatically distributes alerting rules to the monitored resources to solve the trade-off between high accuracy and low network traffic volume without administrative overhead. We evaluate our approach against requirements of today's IT monitoring and compare it to conventional agent-based monitoring approaches.