I have the great pleasure of working with many different engineers, who come from a wide range of backgrounds. Through this continuous process of helping many great people get past problems in their designs, I have learned a number of important things that I tend to refer to as "engineering best practices." One of these practices is, "Understand what to expect and make sure your observations reflect those expectations, BEFORE moving on in your project." In other words, don't allow convenience to be the driving value in your assessment of these situations. Figure them out. If you don't have time, make note of them and make sure there is an appropriate action plan in place to address them when project schedule allows.
We have all experienced nagging problems which are hard to explain and don't seem to impact the core purpose of our system. Sometimes, these observations come at times where deeper investigation threatens schedule commitments so there is temptation to fix them "as easy as possible." While that temptation is very real, taking the time understand the root cause of an unexpected behavior will helps us understand what can influence that behavior (in a positive or negative direction). This leads to better judgments on longer-term risk and if necessary, changes in the design to close these gaps.
Remember, undesirable behaviors are typically there for a reason and can reappear at times that are far less convenient and cause much bigger problems than a few hours in the lab, especially when there is risk of failure, during field operation. If you don't know what to expect, ask! If you don't know what to do, ask!
I will provide some specific examples that relate to my current work, but hope to see feedback from others on this topic as well. Look forward to your feedback.