All of the above. The one issue that is particularly stressing me, is that we have 2 week sprints*. They acknowledge this, accept this, but don't seem to want to adhere to it. My "spies" in the field tell me that the people involved are making remarks about the process, and indirectly us, in a negative light - despite them not even having tried it out yet. They continue to interrupt us mid-sprint and keep taking things the wrong way - such as at estimation time, because of the mess of things due to these same issues, it is difficult for us to acutely estimate, so we have to compensate by allocating a block of time for problem compensation and interuptions. They then think that this 'bucket' of time is at their disposal, despite us agreeing with them that any spare time left in the bucket will be filled with technical debt (of which there is a HUGE amount, because of the way they forced this company to work in the past!)
An example - there was a problem with the product. Our product includes a Linux image and hardware. Some of the hardware was intermittent. We Spiked the investigation of the problem, and even resolved it after we discovered it was not a hardware fault, but a UDEV issue. Immediately they were asking for the new version, despite it being agreed that we do not release until after due diligence for testing etc. which had not been agreed for this sprint - and even if we had, we do not release/deliver anything until the end of the sprint unless we estimate and plan a release mid-sprint... of which, none had happened.
Then they use this as "ammunition" to complain about the process.

Other times they just don't look after their backlog properly, and keep asking us to swap user stories in the current sprint for stories we'd not heard of etc. etc.
*What is a sprint? We estimate and commit to finishing a set amount of work every 2 weeks. Once we begin these two weeks, it is uninterrupted (save for an 'emergency' in which case the sprint is terminated and we begin a new one - i.e. we never change the current sprint, we can only terminate it) and we endeavour to complete this work. Sometimes it doesn't all get done, and we notifty at the first possible chance that it may not be completed, and explain why. Other times the work we estimated was over-estimated and we can pull in an extra story if we are 90% confident it will be completed in time.