diff --git a/modules/policy/README.rst b/modules/policy/README.rst
index d8f95d847d2a296b4ee1b6d80e7b6ccdd5327f0a..0d37b1f6c9c8fb20a0aaf8fe7e357410e76a1f9c 100644
--- a/modules/policy/README.rst
+++ b/modules/policy/README.rst
@@ -36,12 +36,21 @@ A *filter* selects which queries will be affected by specified Actions_. There a
 
    Applies the action if query name suffix matches one of suffixes in the table (useful for "is domain in zone" rules).
 
-.. note:: For speed this filter requires domain names in DNS wire format, not textual representation, so each label in the name must be prefixed with its length. Always use convenience function :func:`policy.todnames` for automatic conversion from strings! For example:
-
    .. code-block:: lua
 
       policy.add(policy.suffix(policy.DENY, policy.todnames({'example.com', 'example.net'})))
 
+.. note:: For speed this filter requires domain names in DNS wire format, not textual representation, so each label in the name must be prefixed with its length. Always use convenience function :func:`policy.todnames` for automatic conversion from strings! For example:
+
+.. _IDN:
+
+.. note:: Non-ASCII is not supported.
+
+   Knot Resolver does not provide any convenience support for IDN.
+   Therefore everywhere (all configuration, logs, RPZ files) you need to deal with the
+   `xn\-\- forms <https://en.wikipedia.org/wiki/Internationalized_domain_name#Example_of_IDNA_encoding>`_
+   of domain name labels, instead of directly using unicode characters.
+
 .. function:: domains(action, domain_table)
 
    Like :func:`policy.suffix` match, but the queried name must match exactly, not just its suffix.
@@ -691,6 +700,8 @@ Response policy zones
 
         log_groups({'policy'})
 
+     See also :ref:`non-ASCII support note <IDN>`.
+
 
 .. function:: rpz(action, path, [watch = true])
 
diff --git a/modules/predict/README.rst b/modules/predict/README.rst
index 634c8d45ccfaa0b0ec24cb80c089998337d799e7..966c4ca4e1d73fd8aa2a04f44dc09b04a1781d6d 100644
--- a/modules/predict/README.rst
+++ b/modules/predict/README.rst
@@ -22,7 +22,7 @@ Prediction
 ----------
 
 The predict module can also learn usage patterns and repetitive queries,
-though this mechanism is basically a prototype.
+though this mechanism is a prototype and **not recommended** for use in production or with high traffic.
 
 For example, if it makes a query every day at 18:00,
 the resolver expects that it is needed by that time and prefetches it ahead of time.
@@ -41,13 +41,12 @@ Example configuration
 
 	modules = {
 		predict = {
+                        -- this mode is NOT RECOMMENDED for use in production
 			window = 15, -- 15 minutes sampling window
 			period = 6*(60/15) -- track last 6 hours
 		}
 	}
 
-Defaults are as above: 15 minutes window, 6 hours period.
-
 Exported metrics
 ----------------