SDL Tridion Sites Infrastructure Suggestions

Please find some of the suggestions while setting up SDL Tridion Sites.


Please note the following suggestions will vary for different Customers and based on implementation. Please Contact SDL for specific questions.

ComponentSuggestions
Content Manager
Suggested to setup
Master/Slave setup
Refer to Search system diagram
(outscale)
Link
Uploads to the Core
Service for scaled out
setup
Suggested to setup to store
the images in either shared
folder or S3 bucket.

Configuring a shared network
temporary location for uploads
to the Core Service
Link
Content Porter – Import /
Export service for scaled
out setup
Suggested to configure this
setup
Configuring high-availability
storage for Import Export
service items
Link
Content Delivery
Content Deployer ServiceSuggested to keep more
memory rather than more CPU, to achieve the mass publishing requirement
Recommended to exclude the
binaries from cache, because it
consume the memory quickly


Suggested to keep the state
store database separately
It also helps to improve the
performance of broker database
and syncing of database
between the environments
Suggested to setup load
balancer for deployer and setup
both servers’ configurations as
identical, and store the packages
in common shared folder, eg EFS
And use the load balancer
endpoint with discovery
registration. Turn off the 2nd
deployer service.

In case first deployer service
down or crashed, then just
switch on the 2nd deployer
server it will without any manual
configuration changes.
Content ServiceRecommended to exclude the
binaries from cache, because it
consume the memory quickly
CMS Load BalancerSuggested to keep
minimum 60 mins or higher.
CMS Load balancer endpoint
mandatory to set sticky cookie
to use the Health check protocol
HTTPS AND 443 with status code
200
DatabaseSuggested to set up 2
scheduled jobs
Job 1 – scheduled to run every
4hr with only update statistics
(it can be decided based on how
often content update and
publishing happens)

Job 2 – scheduled to run every
day in midnight with both reindex and update statistics,
ensure there is no much traffic
during the time.
Recommended to run at ideal
time not peak time.
ActiveMQSuggested to setup HA on premises or recommended to use AWS ActiveMQ
For instances storages eg.
HDD/Shared EFS
Recommended to keep the High
I/O performance for better
performance
Session content serviceEnable object caching but
disable the webapp caching and
CIL caching in modelservice
CredentialsRecommended to encrypt all the
credentials in the configs

Leave a Reply

Your email address will not be published. Required fields are marked *