Governance / Quality
Cover data contracts, DQ dashboards, PII policy, deletions, lineage, audits, and severity-driven response paths.
On this page
Data contracts
Treat schema, privacy, freshness, and trust as first-class product surfaces.
Schema contracts, DQ checks, privacy controls, and audit trails are not side notes here; they are operating requirements of the platform.
Quick jumps
Click to jumpMonthly users
200M-250M
Docs use both 200M+ and 250M assumptions
Daily active users
80M
Used for heartbeat and Kafka math
Peak concurrency
15M
Useful for QoE and live ops workloads
Put contracts before dashboards of trust
If schemas and ownership are weak, downstream quality is always reactive.
Ownership
Every event family names an owning team, schema steward, and paging contact before it can publish into the platform.
Schema rules
Required fields, enum compatibility, timestamp semantics, and PII tags are validated before events reach durable topics.
Breaking change path
Breaking changes fork versioned schemas or topics, notify downstream owners, and block deploys until compatibility checks pass.
Make quality visible like a production surface
The page should show DQ as a system with severity, not a checklist hidden in prose.
Investigation path
Duplicate rate
Possible causes: missing Kafka partition, Flink lag, schema change, bad dedupe logic, content dimension join failure, or late-event spike.
Kafka lag
Flink checkpoint health
Bronze volume completeness
Silver DQ results
Gold aggregate publish
Severity cards
P0
Gold table missing or official metric broken
PagerDuty + incident commander
P1
Watch hours dropped 40% or freshness breach
Slack + PagerDuty
P2
Null rate increased or late events spiked
Slack + Jira
P3
Non-critical metadata missing
Backlog / Jira only
Keep governance and privacy operational
PII, deletes, retention, and access controls must feel concrete.
PII classification matrix
Field policy
Direct PII
Never expose raw email in analytics marts
Hash or tokenize before warehouse storage
Restrict raw access to a minimal approved set of services
Must include
Encryption at rest and in transit
RBAC plus row-level and column-level security
Retention policies by data class and layer
GDPR / right-to-delete propagation across lakehouse, warehouse, and feature stores
Audit logs for data access and correction workflows
Regional data residency where required
Use a trust-and-governance framing line
Good answers make contracts, quality, and privacy feel core to the design.
Say This In Interview
Schema contracts, DQ checks, privacy controls, and audit trails are not side notes here; they are operating requirements of the platform.
30-second answer
FastSchema contracts, DQ checks, privacy controls, and audit trails are not side notes here; they are operating requirements of the platform.
2-minute answer
BalancedCover data contracts, DQ dashboards, PII policy, deletions, lineage, audits, and severity-driven response paths. Focus on PRODUCTION, 8 min, Treat schema, privacy, freshness, and trust as first-class product surfaces..
Deep answer
DeepSchema contracts, DQ checks, privacy controls, and audit trails are not side notes here; they are operating requirements of the platform. Then go layer by layer through PRODUCTION, 8 min, Treat schema, privacy, freshness, and trust as first-class product surfaces., including trade-offs, failure handling, and why this shape is right for the consumer SLAs.
Last reviewed June 2026 · By Prasoon Parashar
Numbers are interview assumptions, not real Netflix internal figures.