One of the most prominent physiological responses to stressors is the activation of the hypothalamic-pituitary-adrenocortical (HPA) axis, currently assessed by measuring cortisol concentration in blood plasma. To reduce animal discomfort during sampling, which negatively affects stress biomarkers, current research focuses on noninvasive sampling of media other than blood, for example, saliva. The aim of this study was to assess the suitability of saliva cortisol as a biomarker under different physiological and immunological states in dairy cows. Our objectives were to 1) evaluate the relationship between HPA axis activation and saliva cortisol concentration, 2) investigate effects of some feeding action (as influenced by feed and water consumption) on saliva cortisol concentration, and 3) evaluate the time lag between plasma and saliva cortisol during induced inflammatory conditions by intramammary lipopolysaccharide (LPS) and lipoteichoic acid (LTA) injection. During a specific activation of the HPA axis, a positive correlation ( = 0.75, < 0.0001) between saliva and blood cortisol concentrations was observed with increased ( < 0.01) plasma cortisol concentrations following ACTH administration. Saliva and blood samples were taken before, during, and after drinking, feeding, and ruminating. Only a low correlation between saliva and plasma cortisol concentrations ( = 0.03, = 0.83) but no significant effects of the different feeding actions on saliva cortisol were observed. When compared with basal concentrations, cortisol concentrations in plasma significantly increased during inflammatory responses following LPS and LTA injection. Compared with plasma cortisol, changes in saliva cortisol concentrations occurred at a much lower level within a narrow range and did not necessarily follow changes in plasma. In conclusion, the positive correlation between saliva and plasma cortisol concentration in response to ACTH and inflammation suggests the suitability of saliva cortisol measurement for the HPA axis activation assessment. However, changes in saliva cortisol concentration occur within a very narrow range. Furthermore, not only must variation among individual animals be considered but also variation within the same animal. Only with additional knowledge of the concomitant physiological status of the cow it is possible to correctly evaluate saliva and blood cortisol samples.