Stress Testing

[rstiff]

Initial Stress-Test Approach

The objective of this test is to determine if our AWS architecture setup can handle a lot of work at once without breaking.

Requirements

• full implementation of run_due_checks
• partial implementation of run_check
• partial implementation of monitor creation

Test Parameters

Here is an example of what the test should do.

• Total Monitors: 500
• Interval: 1 minute
• Duration: 48 hours straight

Success Metrics

Here are some questions we should be asking ourselves after doing the initial stress test:

• How precise will the event scheduler?

• Can we concurrently query and run a few hundred monitor checks fast enough?

• Can we detect errors and retry sanely?

• Is it as cheap/scalable as we thought it would be?

• Changes in architecture we may want to consider?