Database administration tasks often involve several actors within the organization. The main actor is the database administrator (DBA) who interacts with other DBAs and with other actors. Thus, sharing knowledge among them is crucial. This knowledge relies on each member's skills, experiences and background, and also on the common activities, conditions, facts, and situations faced during the administration of the database and the setting of the environment. Exchanging knowledge between the DBA and the other actors can only be effective if there is a common interpretive focus and a shared context where they can understand each other and exchange their experiences. The concept of context plays an important role in collaboration, especially in communication, interaction and knowledge sharing.
In addition, and under temporal and financial pressures, this can help a DBA in avoiding misunderstandings and accidents such asdestroying each other's work or imposed system downtime. The paper presents how collaboration improves the contextualization of DBA procedures and the development of the DBA practices. Our work focuses on collaboration between actors with different roles according to an internal observation that is to say from the point of view of the main actor (i.e. DBA) to help him achieve his objective. This is a key element in the building of a support system for DBA acting as a real context-based intelligent assistant system based on a collaborative work. Contextual Graphs are used to illustrate the collaborative approach to database administration tasks.