When a service has more work than a resource can handle immediately, requests may wait. A sustained wait can show up in a queue and in slower responses.
Service Monitoring · Lesson 5
Resource Saturation
Learn to use utilization and queues to decide where a service may be reaching capacity.
Reading Capacity and Demand
Utilization is how much of a resource is being used. A worker pool, for example, might allow four jobs at once. Four busy workers means the pool is fully used at that moment. If another job arrives, it can wait in a queue until a worker becomes free.
Four workers are busy and three jobs are queued. What do you know?
Comparing Over Time
A short burst is different from a queue that grows across several observations. Compare the queue with request latency and traffic in the same window.
Time Busy workers Queued jobs Typical response
09:10 2 / 4 0 80 ms
09:11 4 / 4 5 300 ms
09:12 4 / 4 14 640 ms
The full worker pool and growing queue suggest that the work is arriving faster than it completes in this example. Check what occupies the workers.
Which observation most strongly suggests continuing pressure?

Checking the Cause
The workers can be busy because requests arrived faster, because a dependency slowed down, or because a recent change made each job take longer. Compare request volume, service time, and dependency observations. An additional signal helps choose a response.
Request volume stayed steady while work began taking longer. What might clarify the backlog?
Lesson complete
You finished Resource Saturation
You can now distinguish a busy resource from a sustained backlog.
Link worker counts to the pool's capacity.
Compare queues with request behavior over time.
Investigate the work that occupies a constrained resource.
Keep your learning progress
Create a free account to save this lesson and continue learning on any device.
Create a free account