Accepted answer
At 30 °C it can show up as either, and which one depends entirely on whether the product still elutes under the main peak. Purity is a ratio of areas, so a degradant only costs purity if the method resolves it. Content is a mass against a standard, so a degradant costs content whenever the parent is consumed — resolved or not. Ordered beta-sheet assembly, effectively irreversible, and its endpoint is opalescence you can see rather than a peak you can integrate. That is why the two measurements are not interchangeable and why an unchanged purity figure after an excursion to 30 °C is weak evidence: the method may simply be integrating the degradant along with the parent and reporting the sum as one peak.
Answer first: the degradation pathways worth knowing are hydrolysis, deamidation, oxidation, aggregation and adsorption, and each has a different trigger and a different mitigation.
Hydrolysis cleaves the backbone, most readily at aspartate-proline and aspartate-glycine sequences, and is acid-catalysed. In a dry solid it barely proceeds at all.
The part that matters: oxidation targets methionine, cysteine and tryptophan, adding sixteen daltons per oxygen. It is catalysed by trace metals and promoted by dissolved oxygen and by light.
Aggregation at air-liquid interfaces is established from surface-tension and particle-count studies and is the basis for anti-agitation handling guidance.
The caveat is that none of these pathways can be seen by looking at a vial, and a clear solution can be substantially degraded.
Cold, dry, dark, still. Those four words cover most of the mitigation.