Why do resolved notifications contain old values?

It often confuses users as to why resolved notifications don't contain updated annotations values. Let's dig into why.

Let's say you had an alert like:

- alert: MyAlert
  expr: foo > 100
  annotations:
    description: 'Foo is {{$value}}'

As the alert is firing, the Alertmanager will receive annotations such as "Foo is 102" and "Foo is 153", and use these in notifications. However an resolved notification from the Alertmanager can't contain an annotation like "Foo is 97".

The reason is that alerts in Prometheus aren't simple thresholds applied to values, they can be far more complex. To allow you this freedom, any sample which results from evaluating the alert expression becomes an alert and the sample including its labels and values are available for use in annotation templating. So if there's multiple foo metrics above 100, you'll get an alert for each of them, each of which will individually have their annotation templates expanded.

When a foo metric drops below 100, that sample will no longer be returned from the alerting expression and the alert is no longer firing. As there's no sample, there's no annotation templates expanded. Prometheus will detect that an alert has gone away, and inform the Alertmanager that the alert with the given labels is resolved.

The Alertmanager knows that the alert is resolved, but neither Prometheus nor the Alertmanager know what the current value of the series that caused the alert to fire is - if there even was exactly one series that was the cause. So when sending resolved notifications, the Alertmanager will use the annotations from the last firing alert it saw.

This is why resolved notifications contain old values, and is one of several things to be wary of with resolved notifications. To see the current value, a better way is to link to a dashboard from your notifications.

内容概要:本文详细介绍了基于滑模控制(SMC)和H∞控制相结合的方法应用于永磁直线同步电机(PMLSM)的鲁棒控制。首先阐述了PMLSM的基本数学模型及其物理意义,包括d-q坐标系下的电压方程和运动方程。随后解释了滑模控制的工作原理,重点在于如何构建滑模面并确保系统沿此面稳定运行。接着讨论了H∞控制的目标——即使在不确定条件下也能保持良好的性能表现。文中还提供了具体的Matlab代码实例,展示了如何利用Matlab控制系统工具箱进行H∞控制器的设计。最后给出了一段完整的Matlab程序框架,演示了两种控制方法的具体实现方式。 适合人群:从事电机控制领域的研究人员和技术人员,尤其是那些想要深入了解滑模控制和H∞控制理论及其在实际工程中应用的人士。 使用场景及目标:适用于需要提高永磁直线同步电机控制系统抗干扰能力和鲁棒性的场合,如工业自动化生产线、精密加工设备等。通过学习本篇文章提供的理论知识和编程技巧,读者能够掌握这两种先进控制策略的应用方法,从而提升自身解决复杂控制问题的能力。 其他说明:文中所涉及的内容不仅限于理论讲解,还包括了大量的实战经验分享,有助于读者快速上手并在实践中不断改进自己的设计方案。同时鼓励读者积极尝试不同的参数配置,以便找到最适合特定应用场景的最佳解决方案。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包

打赏作者

旷野九思

你的鼓励将是我创作的最大动力

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付:¥1
获取中
扫码支付

您的余额不足,请更换扫码支付或充值

打赏作者

实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值