8.11. Conditional access policies

Added in version 3.14.

The scope conditional_access configures how Conditional Access behaves towards the end user. The rules that decide whether a request is refused are not in this scope - they are the conditional access policies of their own, see Conditional access policies.

8.11.1. show_default_ca_error_message

type: bool

Tell the user why conditional access turned their request away, using privacyIDEA’s default wording for the action that did it - for example Your account is temporarily locked. Please try again in about 10 minute(s). for a timed lock, or Access has been denied. for a DENY.

Without this policy a rejection says only what the triggering stage was given its own error message to say, and the generic Authentication failed. where none was written. A stage’s own error message always takes precedence over this policy.

Because the wording is per action, a stage that restricts more than one thing - the user and the source address - is described by one sentence per restriction in force, most severe first. There is wording only for the actions that turn a request away, so a stage that locks the user and emails them is described by the lock alone: the email is a one-off event that no lock records, and nothing is left to report it on a later request.

The default wording is not stored on the lock or the block it describes - this policy is matched on every request - so switching it on or off immediately changes what the locks and blocks already in force say, with no re-locking involved. It only fills in a missing message and never replaces one an administrator wrote on a stage.

Note

This policy is independent of hide_specific_error_message (authentication scope) and of no_detail_on_fail (authorization scope). Neither of those masks conditional access wording - see Interaction with the masking policies.

Warning

Enabling this makes a lock, a block or a denial distinguishable from a wrong password. That is a deliberate trade: it tells a legitimate user whether to wait or to call the help desk, and it tells an attacker that they found a real account and hit a limit.

See Error messages for the whole picture: what each endpoint says, which restriction is described, and how the wording of a lock is stored.