The question:
A facility has three machines and three mechanics. Machines break down at a rate of $1$ per $24$ hours. Breakdown times are exponentially distributed. The time it takes a mechanic to fix a machine is exponentially distributed with mean $6$ hours. Only one mechanic can work on a failed machine at any given time. Let $X_t$ be the number of machines working at time $t$. Find the long-term probability that all machines are working.
My issue is with finding the rate matrix $Q$. Is it as simple as I think? Could I just say it is
$$ Q= \begin{pmatrix} -\frac16&\frac16&0&0\\ \frac1{24}&-\frac5{24}&\frac16&0\\ 0&\frac1{24}&-\frac{5}{24}&\frac16\\ 0&0&\frac1{24}&-\frac{1}{24} \end{pmatrix}\ \ ? $$
My only problem is figuring $Q$ out, please do not provide an answer with the long-term probability. I'd prefer to do that myself. I just want to be sure that this $Q$ is correct, but I've a feeling it isn't considering I never used the fact that they are exponentially distributed. If this isn't the correct approach, would it be better to find the transition probability matrix function $P(t)$ first and then calculate $Q$ from there? If so, any tips on how I'd find this? Thanks.