Concept questionMedium~7 minData & APIs / 11 of 33

Interview question

How would you explain database indexes when diagnosing a slow API endpoint?

How do indexes improve API query performance?

Connects API query shape to index order, selectivity, execution plans, write cost, and safe production rollout.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

An index gives the database an ordered access path so it can find matching rows without scanning the entire table. I start from the endpoint's actual filters, joins, and ordering rather than indexing columns in isolation. For an order list filtered by CustomerId and ordered by CreatedAt, a composite index beginning with those columns can support both lookup and ordering. The cost is extra storage and additional maintenance on writes.

Full answer below

Practice focus

Index seeks and scans, composite index order, selectivity, covering columns, execution plans, pagination, and write-path trade-offs.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Why this question matters
An index is useful only when its key order supports a real predicate and ordering pattern. The same index also consumes storage and adds work to every affected insert, update, and delete.
An index gives the database an ordered access path so it can find matching rows without scanning the entire table. I start from the endpoint's actual filters, joins, and ordering rather than indexing columns in isolation. For an order list filtered by CustomerId and ordered by CreatedAt, a composite index beginning with those columns can support both lookup and ordering. The cost is extra storage and additional maintenance on writes.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
I confirm the need with the generated SQL and execution plan. Column order matters: equality predicates commonly lead, followed by range or sort columns, and a unique tie-breaker can support stable pagination. Included columns may avoid lookups for a hot projection, but making every index covering creates a large and expensive write surface. Low selectivity, stale statistics, parameter skew, or a predicate that wraps the indexed column can still make the optimizer choose a scan.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
In production I design the API contract and index together. A tenant order feed might use (TenantId, Status, CreatedAt DESC, Id DESC) because those columns match its security filter and cursor order, while the response projection determines whether a few non-key columns are worth including. I test with realistic cardinality and representative parameter values, not an empty development table. The rollout also matters: a large index build can lock, consume log space, or compete with normal traffic, so I use the database's safe online or staged migration options where available. After release I monitor latency, rows examined, plan changes, write throughput, and index usage, then remove indexes whose ongoing cost is not justified.

Coding practice

Practice the problem like a real interview: restate, reason, implement, and test.

ExampleCandidatePractical exampleWalk through sample input and output.Open details

Details

For GET /orders?customerId=42&status=open&sort=createdAt, a useful index might start with tenant/customer filters and include status and created date depending on the database and query plan.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Treating indexes for API query performance as a syntax detail instead of a production behavior.
Not checking generated SQL, query plans, or actual data volume.
Testing only with tiny local data and assuming production will behave the same.
Fixing symptoms with caching, timeouts, or retries before understanding the data-access cause.

Red flagsCandidateSigns the approach is fragileSignals that an interviewer may treat as risky.Open details

Details

Cannot explain how indexes for API query performance affects API behavior.
Gives only a definition with no practical example.
Ignores production failure modes.
Does not know how to verify the behavior.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

How do you decide whether an index is needed?
What is selectivity?
How can indexes hurt writes?
How do indexes support sorting and pagination?

Continue the section

Next in Data & APIs: API Compatibility

Step 12 of 33

Design request and response boundaries for API compatibility

Opening drill

Concept questionMedium~7 minData & APIs / 11 of 33

Interview question

How would you explain database indexes when diagnosing a slow API endpoint?

How do indexes improve API query performance?

Connects API query shape to index order, selectivity, execution plans, write cost, and safe production rollout.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

Full answer below

Practice focus

Index seeks and scans, composite index order, selectivity, covering columns, execution plans, pagination, and write-path trade-offs.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Why this question matters
An index is useful only when its key order supports a real predicate and ordering pattern. The same index also consumes storage and adds work to every affected insert, update, and delete.
An index gives the database an ordered access path so it can find matching rows without scanning the entire table. I start from the endpoint's actual filters, joins, and ordering rather than indexing columns in isolation. For an order list filtered by CustomerId and ordered by CreatedAt, a composite index beginning with those columns can support both lookup and ordering. The cost is extra storage and additional maintenance on writes.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
I confirm the need with the generated SQL and execution plan. Column order matters: equality predicates commonly lead, followed by range or sort columns, and a unique tie-breaker can support stable pagination. Included columns may avoid lookups for a hot projection, but making every index covering creates a large and expensive write surface. Low selectivity, stale statistics, parameter skew, or a predicate that wraps the indexed column can still make the optimizer choose a scan.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
In production I design the API contract and index together. A tenant order feed might use (TenantId, Status, CreatedAt DESC, Id DESC) because those columns match its security filter and cursor order, while the response projection determines whether a few non-key columns are worth including. I test with realistic cardinality and representative parameter values, not an empty development table. The rollout also matters: a large index build can lock, consume log space, or compete with normal traffic, so I use the database's safe online or staged migration options where available. After release I monitor latency, rows examined, plan changes, write throughput, and index usage, then remove indexes whose ongoing cost is not justified.

Coding practice

Practice the problem like a real interview: restate, reason, implement, and test.

ExampleCandidatePractical exampleWalk through sample input and output.Open details

Details

For GET /orders?customerId=42&status=open&sort=createdAt, a useful index might start with tenant/customer filters and include status and created date depending on the database and query plan.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Treating indexes for API query performance as a syntax detail instead of a production behavior.
Not checking generated SQL, query plans, or actual data volume.
Testing only with tiny local data and assuming production will behave the same.
Fixing symptoms with caching, timeouts, or retries before understanding the data-access cause.

Red flagsCandidateSigns the approach is fragileSignals that an interviewer may treat as risky.Open details

Details

Cannot explain how indexes for API query performance affects API behavior.
Gives only a definition with no practical example.
Ignores production failure modes.
Does not know how to verify the behavior.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

How do you decide whether an index is needed?
What is selectivity?
How can indexes hurt writes?
How do indexes support sorting and pagination?

Continue the section

Next in Data & APIs: API Compatibility

Step 12 of 33

Design request and response boundaries for API compatibility

How would you explain database indexes when diagnosing a slow API endpoint?

Practice focus

Answer path

Why this question matters

Coding practice

Review signals

Follow-up prompts

Continue the section

How would you explain database indexes when diagnosing a slow API endpoint?

Practice focus

Answer path

Why this question matters

Coding practice

Review signals

Follow-up prompts

Continue the section