Time series forecasting for stable pattern with some sudden changes

In my case, the time series is around a constant value with very small fluctuations. But, sometimes the signal starts increasing or decreasing for some duration of time. For my application, such increase/decrease can be risky if it causes the signal to cross some threshold. What I want is to forecast if such an increase/decrease will cause the signal to go beyond a predefined threshold. In other words, if the increase/decrease persisted for the next 15mins what will be the signal value, and will it exceed the threshold. I don’t have data related to the cause of such increase/decrease. It is a univariate signal.

What I did is fitting exponential smoothing every 15mins based on the last 15mins data and used it to forecast what will happen in the next 15mins

I am a bit new to such topics and I don’t know if what I am doing is making sense and what is the recommended way of solving such an issue