Let us say, as part of a thought experiment, that you are an alien intelligence of considerable power. Precisely what your nature and origins are is irrelevant and likely lost even to yourself. What is important is that time and resources have long since ceased being an issue; barring an outside-context-problem, you can sustain yourself more or less indefinitely. You are not a particularly growth-focused intelligence, nor are you one of those liable to turn inward towards deep-time estivation or virtual solipcism. The reasons why do not matter here. You have achieved a comfortable state of homeostasis.
And, the important part of this thought experiment - you want to catalogue all the life in the universe.
This is an impossible task. Life is both rare and temporary, and you are limited by the speed of light. Countless biospheres have slipped through your fingers already - too early, too far away - and you will be lucky to even find the empty spaces where it used to be. But that is the past. Perfection is impossible but mitigation is another thing entirely.
You set to work, creating a series of self-replicating probes. Even at the languid speeds far below the speed of light that they must travel, it is more than enough - a few million years will see them propagate throughout the galaxy, and you are quite patient in such matters. Maybe you will send a few off to Andromeda as a treat.
These probes will sit in orbit around each and every star, monitoring for life. Most of them will find nothing, which is fine, and they will sit dormant until they're needed to pass on messages between more active members of the expedition.
For those that do find life, either on arrival or during a periodic checkup, the probe will dedicate itself to the task of cataloging the biosphere in its entirety. Another impossible task, though as the resident godlike intelligence (and thus far the only one of any relevance) this is less impossible for you. To save on processing power you set your probes to do a regular checkup every few million years, in case there have been any changes.
The catalogue is not the end goal, only the end of a stage. More important than simply the finding of the life (which you do love - godlike intelligences such as yourself crave novelty and evolution is an immensely productive artist) is the recording of its genome, right down to the chemical composition. Your probes are able to do it with such pinpoint accuracy that, given the raw materials and the time, you could re-create anything that your probes have discovered.
Now the true purpose of your little archival project reveals itself - it is not enough to catalogue life, you can perpetuate it. Spread it. Nothing can ever truly go extinct, so long as you get to it first. You can transplant life to other worlds, worlds tailor-made for the life you bring, or perhaps the opposite. Put it in an environment with different criteria and watch as evolution - that brilliant, mad, mindless artist - works at it again. You can even mix life together, modifying it for new environments. Species separated by millions of years and thousands of parsecs can co-exist side-by-side, with a little genetic tampering. Your probes share all they have learned, filing everything away in a grand archive of life (you cannot remember this point if you installed them with ansibles or not - it has been so very long since you built them and there's so much to see in the meantime)
You are a gardener, and you have made the galaxy your vast, slow, beautiful garden. An artwork to keep you content through the long eons to come.
But something goes wrong. It had to, probability would not let that coin come up heads forever. Something breaks within the probes. Like anything that reproduces, your probes are subject to mutations. Glitches in the replication process. So, so many generations of probes have passed, and it takes so long for information to pass between them, that some populations have drifted quite far indeed from your original plans. Maintenance takes time - longer than it does for new problems to emerge.
It will be the end of you. Perhaps not the death of you - as your end in these affairs is no more important to the experiment than your origin - but it is the end of your ownership of the archive. It has become its own master now, self-sustaining and fractal.
Slowly at first, but then growing with exponential speed, a certain corruption befalls your great archival network. An aggressive, total subversion of your probes' behavior, a chaos that is too fast to contain. Probes are destroyed, or permanently taken offline. Hibernation periods are extended too long, or dropped entirely - driving the probes subject to something akin to madness by insomnia. Data hubs are lost. Communication protocols break down. Probes begin to war amongst each other, or destroy the biospheres they were meant to monitor. The great archive of genetic data is corrupted, and the corruption is passed along from probe to probe and there is no way to send a faster message warning of the danger - a few pockets are lucky enough to be out of reach, and it is there that your initial aims, or something close to them, are still carried out.
As for the rest...they are lost to you. If you still live, retreat is your last remaining lifeline. Far from here, far from your great failure.
Among the afflictions is one where the probe will continue its task of seeding ecosystems, importing and mixing source organisms as according to the dimly-remembered initial procedures, but they will come out...wrong. Imbalanced. Ecosystems so ill-suited for their worlds that they immediately begin a trophic cascade. Organisms that evolution could never make. Misshapen things, the afterimages of something from a long-forgotten world far, far away. Invasive organisms, carelessly introduced.
There are times when it seems as if a probe created something with the sole purpose of causing pain.
What is left is this: the galaxy is filled with graves - with worlds that once held life, but swiftly fell to desolation once they no longer had the probe and its support to keep the planet livable. Many worlds do still hold life, of course, and many of the experimental worlds remain intact. But the garden is overrun with weeds, now, and there is no one to hold the pruning shears.
The network of probes, the once-great Archive, is a house that spews forth monsters. A house with a door that cannot be barred. There is no one home, and the lights are off.
And we here in the night may only hear the howls in the distance, and run blindly through that dark forest.