The classical three-term description of Brahman: being, consciousness, bliss. Not three qualities Brahman has, the way a table has colour and weight. Three aspects of one undivided reality — each pointing at the same ground from a different angle.
These three terms appear together as a single compound — saccidānanda — in the post-Upanishadic tradition. In the Upanishads themselves, the three appear separately: Taittirīya 2.1 defines Brahman as satyaṃ jñānam anantam (truth, knowledge, infinite); Chāndogya 6.8 uses sat (pure being) throughout; the Ānandavallī of Taittirīya 2 traces the hierarchy of joy from human to Brahman-bliss.
Sat — Brahman is not a thing that exists, the way a table exists. Brahman is existence itself. The ground without which no thing could exist at all. When everything else is negated, what remains is Sat — pure being that cannot be negated because even negation requires it.
Cit — Brahman is not a thing that is conscious, the way a person is conscious. Brahman is consciousness itself — the knowing-ground without which no knowing could occur. Pure awareness before it divides into subject and object.
Ānanda — Brahman is not a thing that is happy. Brahman is fullness itself — the completeness that lacks nothing. Human bliss is always about gaining or not-losing something. Brahman-ānanda is the fullness of what has never lacked anything and can never lose anything.
The Taittirīya's definition: Brahman is truth (satya), knowledge (jñāna), infinite (ananta). These are not predicates of Brahman in the ordinary sense — if Brahman is said to be knowing, knowing would be an attribute added to Brahman, which contradicts Brahman's indivisibility. The Advaita reading: these are exclusionary qualifiers. Satyam excludes the unreal (Brahman is not māyā). Jñānam excludes the insentient (Brahman is not matter). Anantam excludes the limited (Brahman is not any finite entity). Together they point at what is left when all finite, insentient, and unreal objects have been excluded.
The compound saccidānanda first appears explicitly in the later Advaita tradition (Vivekacūḍāmaṇi, attributed to Śaṅkara, uses the compound). It synthesises three different Upanishadic entry-points: Sat from the Chāndogya's cosmology, Cit from the Aitareya's Prajñānam Brahma, Ānanda from the Taittirīya's Ānandavallī. Each entry-point addresses a different question: What is real? (Sat.) What knows? (Cit.) What is the nature of fulfilment? (Ānanda.) The three together make the point that these are not three different answers about three different things. They are three aspects of one reality.
Śaṅkara's technical distinction: sat, cit, and ānanda are not viśeṣaṇas (qualifying attributes) of Brahman — they are svarūpa-lakṣaṇas (intrinsic pointers, indicators of essential nature). The difference matters: an attribute is separable from its subject (a table can be red or not-red). An intrinsic indicator is inseparable — it does not qualify Brahman but indicates what Brahman is. Satyam does not mean Brahman has the property of truth; it means truth is what Brahman is. Similarly for jñānam and anantam.
Taittirīya Upaniṣad 2.5–8 presents the most extended treatment of ānanda: a progression from human bliss upward through the bliss of gandharvas, gods, Indra, Bṛhaspati, Prajāpati, to Brahman's bliss, each a hundredfold increase over the previous. The final statement: the bliss of Brahman is the measure against which all bliss is measured. The Advaita reading: the hierarchy does not describe degrees of the same thing (as if Brahman's bliss is just a lot more pleasant than human bliss). It describes the progressive dissolution of the sense of limited self that obscures the ānanda that is already and always the ground. Liberation is not gaining bliss — it is the falling away of what was obscuring the fullness that was never absent.