Question 1

What are SLAs, SLOs and SLIs?

Accepted Answer

Three related things at different levels: an **indicator** is what you measure, an **objective** is the target you set internally, and an **agreement** is the contractual promise with consequences.

**Service Level Indicator (SLI).** A **measurement** of some aspect of service. Request success rate, latency at a given percentile, availability, throughput, or freshness of data.

**Getting the indicator right is most of the work.** It should measure what users actually experience — availability measured at the load balancer can look excellent while users receive errors from a downstream dependency. **Measure from as close to the user as possible.**

**Percentiles rather than averages.** An average latency conceals a badly affected minority; **p99** tells you what the worst-served users experience, and is what people notice.

**Service Level Objective (SLO).** A **target** for an indicator over a window — for example, 99.9% of requests succeeding over 30 days. Internal, set by the team, and the number that actually drives decisions.

**Why it should not be 100%.** Perfect reliability is unattainable, and pursuing it is enormously expensive with diminishing returns. **Users cannot distinguish 99.99% from 100%**, because their own network and device are less reliable than that. Choosing a target below perfect is a deliberate decision about how much reliability is worth.

**Error budget** is the consequence, and it is the genuinely useful idea. **100% minus the SLO is the amount of unreliability you have permitted.** A 99.9% objective over 30 days allows roughly 43 minutes of failure.

**What the budget does:** it converts reliability from an argument into a measurement. Budget remaining means shipping can continue; budget exhausted means stopping feature work and addressing reliability. **It replaces a perpetual disagreement between development and operations with a shared number**, which is its real function.

**Service Level Agreement (SLA).** A **contract** with a customer, specifying a level of service and the **consequences** of failing to meet it — usually service credits.

**SLAs should be looser than SLOs**, deliberately. You want to breach your internal objective and act on it long before you breach a contractual commitment.

**Availability figures are meaningless without the measurement definition**, which is where providers differ substantially.

Question 2

What is a message queue and when do you need one?

Accepted Answer

A message queue sits between a producer of work and a consumer of it, holding messages until they can be processed. It converts a synchronous call into an asynchronous one. What it gives you:

Question 3

What are the best practices for microservices architecture?

Accepted Answer

Implementing best practices for microservices architecture enhances scalability and maintainability. Key practices include API-first design, resilience, and continuous deployment.

Question 4

What is the difference between scaling up and scaling out?

Accepted Answer

Scaling up means a bigger machine; scaling out means more machines. They fail at different points and impose completely different demands on your software. Scaling up — vertical. More cores, more

Question 5

What is observability, and how is it different from monitoring?

Accepted Answer

Monitoring tells you whether the things you anticipated are behaving. Observability is the property of being able to answer questions you did not anticipate, from the data the system already emits.

Question 6

How do you find a performance bottleneck?

Accepted Answer

By measuring before changing anything, because intuition about where time goes is reliably wrong — and the great majority of wasted optimisation effort comes from skipping that step. The order that

Question 7

What does being on-call actually involve?

Accepted Answer

Being responsible for responding to production incidents outside working hours — and it is one of the more consequential aspects of software work for people's lives, and one of the least well

Related Questions