When you go the extra mile to solve a problem,you're displaying heroic behaviour. The next time the problem occurs, you once again jump in to solve the problem, in addition to the other work you need to do. Over time you keep jumping in. In many organisations, that will be appreciated, but there is a problem with it. Because you jumping in means that you either work more hours, or spend less time on the other things that are expected of you. Both are far from an ideal solution. Hero behaviour doesn't scale, because in the end you will burn out.
Hero behaviour means that the rest of the organisation does not find out about the problem. And does not get a chance to learn and improve. In the short term the problem is solved, in the long term nothing changes. The scalable and long term solution to the problem might take more time initially but will save time over the long term.
Create a problem by solving a problem
The problem you are solving with your hero behaviour provides data the organisation needs to decide where improvements are needed. If the product team never hears about an issue because you are solving it without their knowledge, they will never know they need to improve it. Or if only part of all the cases reach them, because only part of all cases come in via the official channels, their decision about what to prioritise will be based on skewed data. Letting the fire burn to ensure the severity of the problem is clear to all involved will be the best in the long term.
Individual action vs hero behaviour
In Holacracyyou are expected to take individual action to put out the fire if there is no role that is accountable for putting out the fire, but then to use this situation as a tension that needs to be governed into a role or accountability. It's ok, and even expected, to be a hero, but only once. Then the organisation should learn, adapt, and provide the long term solution to the problem.
An example of hero behaviour
At Springest, I realized that there was no role responsible for checking and publishing user reviews. I found 100s of unpublished reviews. So at night, after work, I spent some hours checking and publishing reviews. And I kept checking every day after work to see if there were any new reviews that needed to be published.
After some weeks I proudly shared that I had been doing all this work, expecting praise. I got critique instead. Why had I shown this hero behaviour? We could have easily fixed this problem if we had known about it. And I could have spent my time on something more useful. The next day a developer made it so user reviews were automatically published (if they met certain conditions) and we governed a role to check and publish the remaining user reviews. I got my evenings back, we automated a process that made us more scalable, and the organisation was able to adapt to reality.