One Missing Observation Then Update

A local-level model satisfies $x_t=x_{t-1}+w_t$ with $w_t\sim N(0,3)$, and observations have noise variance 5. After the last filtered state $N(-1,9)$, there are 1 consecutive missing observations. Then you observe a new value $y=2$. Compute the variance just before the new observation and the updated mean/variance after processing it.