summary refs log tree commit diff
path: root/README.rst
diff options
context:
space:
mode:
authorRichard van der Hoff <1389908+richvdh@users.noreply.github.com>2019-06-18 23:29:58 +0100
committerGitHub <noreply@github.com>2019-06-18 23:29:58 +0100
commit8fcd2ca907742b5087cec6058161d2effa55bcb3 (patch)
treea1155e35006f7003ca6ddabf7ba3c4a9d5d21fd9 /README.rst
parentAdd some logging to 3pid invite sig verification (#5015) (diff)
parentREADME.rst: fix header level (diff)
downloadsynapse-8fcd2ca907742b5087cec6058161d2effa55bcb3.tar.xz
Merge pull request #4276 from Ralith/performance-advice
Improve advice regarding poor performance
Diffstat (limited to 'README.rst')
-rw-r--r--README.rst32
1 files changed, 25 insertions, 7 deletions
diff --git a/README.rst b/README.rst
index 5409f0c563..13e11a5773 100644
--- a/README.rst
+++ b/README.rst
@@ -340,8 +340,11 @@ log lines and looking for any 'Processed request' lines which take more than
 a few seconds to execute. Please let us know at #synapse:matrix.org if
 you see this failure mode so we can help debug it, however.
 
-Help!! Synapse eats all my RAM!
--------------------------------
+Help!! Synapse is slow and eats all my RAM/CPU!
+-----------------------------------------------
+
+First, ensure you are running the latest version of Synapse, using Python 3
+with a PostgreSQL database.
 
 Synapse's architecture is quite RAM hungry currently - we deliberately
 cache a lot of recent room data and metadata in RAM in order to speed up
@@ -352,14 +355,29 @@ variable. The default is 0.5, which can be decreased to reduce RAM usage
 in memory constrained enviroments, or increased if performance starts to
 degrade.
 
+However, degraded performance due to a low cache factor, common on
+machines with slow disks, often leads to explosions in memory use due
+backlogged requests. In this case, reducing the cache factor will make
+things worse. Instead, try increasing it drastically. 2.0 is a good
+starting value.
+
 Using `libjemalloc <http://jemalloc.net/>`_ can also yield a significant
-improvement in overall amount, and especially in terms of giving back RAM
-to the OS. To use it, the library must simply be put in the LD_PRELOAD
-environment variable when launching Synapse. On Debian, this can be done
-by installing the ``libjemalloc1`` package and adding this line to
-``/etc/default/matrix-synapse``::
+improvement in overall memory use, and especially in terms of giving back
+RAM to the OS. To use it, the library must simply be put in the
+LD_PRELOAD environment variable when launching Synapse. On Debian, this
+can be done by installing the ``libjemalloc1`` package and adding this
+line to ``/etc/default/matrix-synapse``::
 
     LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libjemalloc.so.1
 
 This can make a significant difference on Python 2.7 - it's unclear how
 much of an improvement it provides on Python 3.x.
+
+If you're encountering high CPU use by the Synapse process itself, you
+may be affected by a bug with presence tracking that leads to a
+massive excess of outgoing federation requests (see `discussion
+<https://github.com/matrix-org/synapse/issues/3971>`_). If metrics
+indicate that your server is also issuing far more outgoing federation
+requests than can be accounted for by your users' activity, this is a
+likely cause. The misbehavior can be worked around by setting
+``use_presence: false`` in the Synapse config file.