More in Monitoring & SRE — page 6
Essential tags for a request latency metric
WHAT IT TESTS: Choosing useful, bounded metric labels. OUTLINE: Add method, route template, status code, and maybe service or region; they enable slicing while staying low-cardinality. RED FLAG: Adding user IDs or raw URLs, exploding cardinality.
Structured vs unstructured logging
WHAT IT TESTS: Why log format matters at scale. OUTLINE: Unstructured logs are free-text lines; structured logs are machine-readable key-value or JSON, enabling reliable parsing, filtering, and aggregation.
The three pillars of observability
WHAT IT TESTS: Knowing each telemetry type's strength. OUTLINE: Metrics show what and when via cheap aggregates, logs give detailed per-event context, traces show where across services.
Fixing watermelon SLOs that hide unhappy users
WHAT IT TESTS: Whether SLIs truly track user experience. OUTLINE: Green-outside, red-inside SLOs come from server-side measurement, coarse aggregation, wrong events, and lax thresholds; fix by measuring at the client and using meaningful SLIs.
Launching a risky feature with no error budget left
WHAT IT TESTS: Using error budgets for negotiation, not dogma. OUTLINE: Bring burn history, what consumed the budget, blast radius, and rollback safety; offer mitigations like flags and canaries. RED FLAG: A flat veto or a rubber-stamp yes with no data.
Setting SLOs across a dependency chain
WHAT IT TESTS: Reliability math across dependencies. OUTLINE: Serial dependencies multiply, so each backend needs a tighter SLO than the target; mitigate with redundancy, caching, and graceful degradation.
Handling planned maintenance in SLOs and error budgets
WHAT IT TESTS: Honest accounting of planned downtime. OUTLINE: Either charge maintenance to the error budget or formally exclude it via a maintenance window; each trades honesty against fairness.
Defining SLOs for a new critical service
WHAT IT TESTS: Cross-functional SLO design. OUTLINE: Start from user journeys, pick SLIs, involve product, engineering, and business stakeholders, set realistic targets iteratively.
Diagnosing a healthy p50 but breaching p99
WHAT IT TESTS: Understanding tail latency. OUTLINE: One percent of requests are slow, hurting power users and fan-out calls; investigate GC, locks, contention, cold caches, retries. RED FLAG: Dismissing it because the average looks fine.
Proposing availability and latency SLIs for an auth API
WHAT IT TESTS: Designing measurable, user-centric SLIs. OUTLINE: Availability as the ratio of successful valid requests; latency as the fraction served under a threshold; measure at the edge from the user's view.
What is an error budget?
WHAT IT TESTS: Linking reliability targets to delivery decisions. OUTLINE: Error budget is one minus the SLO, the allowed unreliability; it gates feature velocity versus reliability work.
Relationship between SLI, SLO, and SLA
WHAT IT TESTS: Core reliability vocabulary. OUTLINE: SLI is the measured metric, SLO is the internal target for that metric, SLA is the external contract with consequences. RED FLAG: Treating the terms as synonyms or making the SLO looser than the SLA.
Symptom-based vs cause-based alerting
WHAT IT TESTS: Alerting philosophy and pager hygiene. OUTLINE: Symptom alerts fire on user-visible impact, cause alerts on internal conditions; page on symptoms to cut noise and catch unforeseen failures.
Head-based vs tail-based trace sampling
WHAT IT TESTS: Tradeoffs in trace sampling. OUTLINE: Head-based decides at trace start cheaply but blindly; tail-based waits for the full trace to keep errors and slow requests. RED FLAG: Claiming head-based can preferentially keep failed traces.
What is high-cardinality data in Prometheus?
WHAT IT TESTS: Understanding time-series storage cost. OUTLINE: Cardinality is the count of unique label combinations; each is a separate series; explosion blows up memory and query cost.
Black-box vs white-box monitoring for legacy apps
WHAT IT TESTS: Pragmatic monitoring strategy. OUTLINE: Black-box probes from outside, white-box reads internal state; start black-box for fast user-facing signal. RED FLAG: Choosing white-box first on an uninstrumented system or ignoring black-box blind spots.
How does distributed tracing work?
WHAT IT TESTS: Whether you understand request correlation across services. OUTLINE: Trace ID ties one request together, span IDs are individual operations, context propagates via headers. RED FLAG: Conflating traces with logs or omitting propagation entirely.
Structured vs unstructured logging: why it matters
WHAT IT TESTS: understanding of log formats and their impact on observability at scale. OUTLINE: unstructured logs are free-text lines hard to parse; structured logs are machine-readable key-value or JSON records; structure enables reliable querying…
Why use latency percentiles over the average?
WHAT IT TESTS: understanding of latency distributions and tail behavior. OUTLINE: averages hide the tail and are skewed by outliers, so most users can suffer while the mean looks fine; use percentiles; prioritize p99 when tail users are high-value or fan-out…
What are the Four Golden Signals?
WHAT IT TESTS: knowledge of the core user-facing monitoring signals. OUTLINE: name latency, traffic, errors, and saturation; explain each briefly; describe measuring latency as a distribution separating success from failure.