You use Spinnaker to deploy your application and have created a canary deployment stage in the pipeline. Your application has an in-memory cache that loads objects at start time. You want to automate

Question

Options

• ACompare the canary with a new deployment of the current production version.
• BCompare the canary with a new deployment of the previous production version.
• CCompare the canary with the existing deployment of the current production version.
• DCompare the canary with the average performance of a sliding window of previous production

• A
  85% (23)
• B
  4% (1)
• C
  7% (2)
• D
  4% (1)

Explanation

Because the application uses an in-memory cache loaded at start time, the canary and its baseline comparison target must be on equal footing - both freshly started so both have the same warm-up state. Comparing the canary against a newly deployed instance of the current production version (A) ensures both are starting cold, giving a fair apples-to-apples comparison. Option C (existing production deployment) is wrong because the existing instance's cache is already warmed and its performance may differ due to runtime state. Option B (previous version) would detect regressions, not equivalence. Option D (sliding window average) does not account for cache warm-up bias.

No community discussion yet for this question.