If a source is reliable, then what it tells us is likely to be accurate and very close to the truth of what actually happened. In order to determine the extent of a source's reliability, it is easier to break down our investigation:
-Firstly, WHO wrote the source? Are they likely to be very knowledgable about the event itself - for example, a source about Henry VIII written by one of his closest advisors may be likely to be more valid that one written by a Lord that lived far across the country from the King. Therefore, was the creator of the source close enough to the event/issue to give an accurate account?
-WHEN was the source written? Is it a primary source (produced at the time), or a secondary source? If it is a primary source, its credibility is likely to be strengthened, as a first-hand account will be less ruined by memory, or distorted by gossip, or by people passing stories over generations.
-WHY was the source created - what was the aim or purpose of it being created? In order to answer this part, you need to take into account the audience of the source, and what the source's creator was trying to achieve. For example, if a political leader was trying to increase their popularity and power, they might give a speech to the public that exaggerated their own achievements, and omitted details that might worry the public, such as likely food shortages in the near future.
Once you have assessed whether each of these details makes the source more or less reliable, you need to come to a judgement. Is it very, partly, or not reliable at all? In order to back up your argument on this, it may be useful to compare WHAT the source is saying, to other sources about that event. Do they say similar things, or are there significant differences? Accounts whose details coincide are far more likely to be reliable.