1: \begin{abstract}

2: 

3: 

4: 

5: 

6: 

7: Integrable many-body systems are characterized by a complete set of preserved actions. Close to an integrable limit, a {\it non-integrable} perturbation creates a coupling network in action space which can be short- or long-ranged. 

8: We analyze the dynamics of  observables which become the conserved actions in the integrable limit.

9: We compute distributions of their finite time averages and obtain the ergodization time scale $T_E$ on which these distributions converge to $\delta$-distributions.

10: We relate $T_E$ to the statistics of fluctuation times of the observables, which acquire fat-tailed distributions 

11: with standard deviations $\sigma_\tau^+$ dominating the means $\mu_\tau^+$ and establish that $T_E \sim (\sigma_\tau^+)^2/\mu_\tau^+$.  The Lyapunov time $T_{\Lambda}$ (the inverse of the largest Lyapunov exponent) is then compared to the above time scales.

12: We use a simple Klein-Gordon chain to emulate long- and short-range coupling networks by tuning its energy density.

13: For long-range coupling networks $T_{\Lambda}\approx \sigma_\tau^+$, 

14: which indicates that the Lyapunov time sets the ergodization time, with chaos quickly diffusing through the coupling network.

15: For short-range coupling networks we observe a {\it dynamical glass}, where  

16: $T_E$ grows dramatically by many orders of magnitude and greatly exceeds  the Lyapunov time, which satisfies $T_{\Lambda} \lesssim \mu_\tau^+$. 

17: This effect arises from the formation of highly fragmented inhomogeneous distributions of chaotic groups of actions, separated by growing volumes of non-chaotic regions. These structures persist up to the ergodization time  $T_E$.

18: 

19: 

20: 

21: 

22: 

23: 

24: 

25: \end{abstract}

26: