Management of 
SysAdmin/DevOps
teams
Marian Marinov <mm@yuhu.biz>
Who Am I?
Marian Marinov <mm@yuhu.biz>
Who Am I?
➢ System Administrator since 1996
➢ Head of DevOps at Web Hosting 
Canada
➢ Managing System Administrators for 
more then 15 years
Marian Marinov <mm@yuhu.biz>
Why  this  subject?
Marian Marinov <mm@yuhu.biz>
Management methodologies
Management methodologies
➢ Agile
➢ Waterfall(Water flaw)
➢ DevOps
Mainly for developers
Dev projects
➢After delivery to customers, there is mainly
maintenance.
➢ Maintenance depends on the quality of the
product
Defining the work of 
sysadmin/devops project
➢ Needs to setup the infrastructure
➢nice if:
➢ you have to setup just one DB
➢ spin up your AWS instance and sync the files
➢ very complicated if you have to setup
➢an HA environment
➢cluster
Defining the work of 
sysadmin/devops project
➢ Monitoring
➢ will this be monitored
➢ how much overhead are we adding to the team
➢ new projects come in - they only add to the amount
of current projects
Executing the work
➢ The difference with Devs is that Sysadmins are
always in maintenance mode
➢ security updates (Software, Firewall, IDS/IPS)
➢ feature updates
➢ client software updates
➢ authentication/authorization management
➢ configuration changes
Time management
➢ Most of the time you should plan to have
available resource
➢ problems may happen every day
➢ IT work is offloaded to them, as they know how
to do it
➢ like setup that printer
➢ fix the VPN for that employee
Documentation
➢ Skilled admins can sometimes work without or
with poor documentation
➢ Junior admins ALWAYS require GOOD and
tested documentation
➢ Makeing sure your admin team keeps up with
the documentation guarantees easier recovery
or faster maintenance
Siloed people
➢In many companies we have people that are
the most knowledgeable or even the only
person that can do specific tasks
➢ either document that
➢ or train more people to do it
Similarities with Devs
➢ You have to know your infrastructure
➢ machines, software, network
➢ As the Devs know theirs
➢ framework, databases, worflow
Takeaways
➢ Separate the work week into a few categories
➢ Monitoring
➢ Maintenance
➢ New deployments or upgrades
➢ Long term goal
Takeaways
➢ It is not mandatory you keep all of the hours
the same each week
➢ It is not important that you have all of them
every week
➢ Just try to accomodate everything wihin the
week
Takeaways
➢ Always plan with 50% chance for disaster. So
keep a second view of the backlog... as if there
is a disaster.
➢ I personally always sacrifice the log term goal
in case of emergencies
➢ In cases where the team is overwhelmed I also
delay deployments and even some times
remove peopel from monitoring.
Marian Marinov <mm@yuhu.biz>
 THANK YOU
Marian Marinov <mm@yuhu.biz>

Managing sysadmins