Heuristic
Order troubleshooting tests by likelihood and risk
Do not reboot the world to answer a question a safe check could settle.
When it fits
- Several tests could be run next and some are disruptive, slow or expensive.
When to avoid it
- Likelihood estimates are judgment calls; do not let a familiar checklist override clear contrary evidence.
Why it matters
For each plausible next test, consider how likely its target explanation is, what the test can distinguish, and what harm or state change the test can cause. Run cheap, low-risk, high-plausibility checks first unless stronger evidence justifies a different order. The sequence itself is part of incident safety.
An example
Check whether the expected endpoint is reachable before enabling verbose production tracing that may worsen latency.
Check your result
You can explain why the chosen test comes before the more invasive alternatives.
Keep this limit in mind
- Likelihood estimates are judgment calls; do not let a familiar checklist override clear contrary evidence.
Connected ideas
Useful withKeep a hypothesis–test–result ledger
Evidence and sources
Google SRE recommends considering obvious causes first and ordering tests with both likelihood and risk to the system in mind.
A low-risk common check should not become ritual if evidence already points strongly elsewhere.
Effective Troubleshooting · Test and Treat; test design considerations