Pega Certified System Architect cheat sheet

A case type defines the case lifecycle and is divided into stages; each stage contains processes, and each process contains steps.

A step can be an assignment, automation, or subprocess; an assignment creates a work item for a user or a work queue.

Routing to a work queue lets eligible users pull work; routing to an operator or worklist targets specific users; skills can filter eligible users.

Service levels define goal and deadline intervals; escalation actions execute when the goal or deadline is exceeded.

Pessimistic locking locks a case when a user opens it for editing; optimistic locking allows concurrent opens and detects version conflicts at save.

A subprocess shape calls another process and can be configured to spin off asynchronously; a synchronous subprocess waits for completion.

A wait shape pauses a flow until a condition is satisfied or a specified event occurs, then continues along its outgoing connector.

A decision shape branches based on a when condition or decision table/tree; a split join shape creates parallel paths and can wait for all or some to complete.

Optional actions and ad hoc processes allow users to launch additional processes from a case outside the main lifecycle path.

Property modes: scalar holds one value, page holds one embedded page, page list holds an ordered list, page group holds an unordered collection keyed by a value.

A data page is a named clipboard page loaded from a data source; its scope can be thread, requestor, or application, controlling reuse.

Data page refresh strategies include reload if older than, reload once per requestor, and reload every time the page is accessed.

A savable data page permits write operations; a data transform maps clipboard values to the data source for persistence.

Data transforms can set, append, remove, and update target pages or lists; they execute in order and can use when conditions.

Declarative expressions, constraints, and on-change rules recalculate automatically when source properties change; forward chaining propagates dependent changes.

A linked property stores a reference to another page; it avoids copying all related fields but requires the referenced page to be available.

Optimistic locking compares a version or update timestamp at save; pessimistic locking obtains an exclusive lock before editing.

Class mapping defines how properties persist to database tables; keys uniquely identify instances and support locking and retrieval.

A harness defines the overall screen for an assignment, such as perform, review, or confirm, and hosts sections; a section groups fields and controls.

A flow action specifies the assignment UI and processing, including the harness or section, buttons, and pre- and post-processing.

Dynamic layouts arrange fields in rows and columns and reflow at responsive breakpoints; repeating dynamic layouts repeat content for page lists.

Visibility conditions use when rules or expressions to show or hide UI elements; disable conditions prevent editing.

Action sets bind events such as click, change, or load to actions such as refresh, set value, or navigate.

A control determines how a property is displayed and edited; field value rules can override labels and formatting.

Skins and themes define visual styling; responsive breakpoints control layout at different viewport widths.

Localization uses field value rules for translated text and locale-specific formats; accessibility requires labels and meaningful alternative text.

A grid displays a page list or page group in rows; a repeating dynamic layout is used for custom repeated content.

An operator ID belongs to an access group; the access group selects the application, default ruleset, and access roles.

Access roles grant privileges; role-based access control checks roles, while attribute-based access control policies evaluate attributes and conditions.

Access control policies can permit or deny access to case data and actions based on conditions; they supplement role-based access.

Authentication verifies identity using password, LDAP, SAML, OAuth, or custom authentication; authorization determines permitted actions.

SAML enables single sign-on through an identity provider; OAuth delegates authorization for APIs and external clients.

Sensitive properties can be encrypted at rest; masking can hide values in the UI while preserving the stored value.

Production rulesets are locked versions; rule changes require a new ruleset version and deployment through the ruleset pipeline.

Auditing records changes in history classes and logs; access logs track user activity for security review.

Password policies define complexity, expiration, and lockout thresholds; account lockout reduces brute-force risk.