Skip to main content

Jenkins Executors

Executors control how many jobs run in parallel in Jenkins.
Misunderstanding executors is a common cause of overloaded agents and flaky builds.

What Is an Executor?​

An executor is a single slot on an agent that can execute one job at a time.

  • Executors belong to agents, not jobs
  • Each running build consumes one executor
  • Executors directly impact CPU, memory, and disk usage

Executors vs Agents​

ConceptPurpose
AgentMachine/container that runs jobs
ExecutorParallel execution slot on an agent

One agent can have multiple executors.

How Executors Work​

  • Job enters queue
  • Controller finds a matching agent
  • Controller checks available executors
  • Executor is allocated
  • Job starts execution

If no executor is free → job waits in queue.

Configuring Executors​

  • Executors are configured per agent
  • Default executor count is often 2
  • Controller also has executors (should be set to 0 in prod)

Best Practice​

  • Set controller executors to 0
  • Use agents for execution

Parallelism Strategies​

  • Higher CPU & memory contention
  • Risk of flaky tests
  • Better isolation
  • Horizontal scaling
  • More stable builds

Executors & Pipelines​

In pipelines:

  • Each stage may consume an executor
  • Parallel stages consume multiple executors simultaneously

Example​

parallel {
  stage('Chrome') { steps { sh 'run-tests.sh' } }
  stage('Edge') { steps { sh 'run-tests.sh' } }
}

This requires 2 free executors.

Common Misconfigurations​

  • Too many executors on a small agent
  • Running heavy jobs in parallel on same agent
  • Not accounting for memory-intensive tests

Monitoring Executors​

Watch for:

  • Long queue times
  • Agent CPU spikes
  • Out-of-memory errors

These usually indicate executor misconfiguration.

Interview Traps​

  • Does each stage need an executor?
    → Yes, when it executes steps

  • Where are executors configured?
    → On agents

  • Should controller have executors?
    → No, not in production