Operating a highly available automation system

A method for operating a program-controlled highly available automation system configured redundantly with a first automation device and a second automation device, for a technical process, wherein one automation device of the first automation device and the second automation device preferentially controls the technical process via peripheral units, and the first automation device and the second automation device mutually monitor for failure of the respective other automation device of the first automation device and the second automation device, wherein a monitoring query from the first automation device to the second automation device, and vice versa, is to be responded to within a monitoring time, wherein for a communication, the first automation device and the second automation device, and the peripheral units are connected with one another via a ring, wherein the first automation device and the second automation device each have a first ring port and a second ring port in order to form the ring, wherein one ring port of the first ring ports and the second ring ports is operated as blocked so that circulating telegrams are prevented in the ring, wherein the first automation device and the second automation device, and the peripheral units also have routing tables, wherein a measuring program is operated in at least in one automation device of the first automation device and the second automation device, the method comprising:

• carrying out a desired ring interruption, the carrying out of the desired ring interruption comprising blocking, by the measuring program, a ring port to provoke a ring reconfiguration that utilizes a reconfiguration time;
• canceling, by the measuring program, the blockage of the ring port blocked by the measuring program after the reconfiguration time has elapsed;
• blocking the ring port that has previously been operated blocked so that circulating telegrams in the ring are prevented again when the ring port is opened by the ring reconfiguration and deleting all routing tables, and as a result of which at least the peripheral units are triggered to learn new network routes;
• measuring runtimes of test telegrams and storing, by the measuring programs, a maximum value of the measured runtimes;
• and
• using the measured maximum value for a dynamic adjustment of the monitoring time.