Yes, the national poll that shows Clinton and Trump in a statistical tie may be credible. This does not mean that the poll should be accurate. People would like to see Trump losing in the polls given his recent comments, however, there are a lot of Republicans who would support him anyway.
Politics can be a minefield for the innocent believer of poll results, but surely we can take such a reputable pollster as Reuters/Ipsos at face value? They gave Trump 11 points behind Clinton, at 34.1 to 45.1 respectively, this month. I don't believe they would risk their reputation on giving spurious statistics, so yes, I do think the poll results are credible.
It is not stated where the poll came from or what region/demographic was involved or how many people were in it. The reason this is important is because the poll would not be reliable if it didn't include a substantial amount of participants or was based on a region more inclined to side with republicans or democrats.
Doing a poll is not credible. It is few people and there is no way to determine why someone is picking either choice. They could be on the fence and just choose one person, or they could just pick someone without paying attention. A simple poll has no credibility in making political predictions.