05Design patternsTime-series metrics ingestion

05 · Reusable pattern

Time-series metrics ingestion

Design append-heavy collection, aggregation windows, retention, downsampling, and cardinality controls.
8 minConcept guideReference-informed · independently authored
01

Lesson spine

What you need to understand.

Time-series ingestion is an append-heavy pipeline whose enemies are unbounded cardinality, retention cost, and expensive wide scans.

01

Collection model

Push fits short-lived or edge producers; pull centralizes discovery and scrape control for service metrics.

02

Write path

Batch samples, append to a WAL, update in-memory structures, and flush immutable time-partitioned columnar blocks.

03

Series identity

A metric name plus labels defines a series. Unbounded user IDs or request IDs create cardinality explosions.

04

Compression and retention

Delta-encode timestamps and values, compact blocks, expire raw data, and retain longer rollups.

05

Downsampling

Aggregate old one-second samples into minute, hour, and day resolutions while preserving the functions queries need.

06

Query and alerting

Partition by series and time, prune blocks, precompute common aggregates, and evaluate alerts over streaming or recent windows.

02

Before the boxes

Frame the decision.

Outcome

What must work

Design append-heavy collection, aggregation windows, retention, downsampling, and cardinality controls.

Scale

What changes the design

Samples/s · active series · labels · retention · compression · query window

Boundary

What owns the truth

Identify the component that commits authoritative state, then separate synchronous confirmation from derived work.

Non-goal

What stays simple

Do not add global coordination, multi-region writes, or a specialized store until a requirement earns the complexity.

03

Architecture map

Trace ownership, not just traffic.

Time-series metrics ingestion · concept mechanism

Walk one representative request across every arrow. Say whether the handoff is synchronous or asynchronous, what identity makes a retry safe, and which step changes authoritative state.

  1. 01

    Agents — Batch samples Define the output contract before moving to the next owner.

  2. 02

    Gateway — Normalizes Define the output contract before moving to the next owner.

  3. 03

    Partitioner — Routes by series and time Define the output contract before moving to the next owner.

  4. 04

    Write log — Buffers ingestion Define the output contract before moving to the next owner.

  5. 05

    Compactor — Builds columnar blocks Define the output contract before moving to the next owner.

  6. 06

    Downsampler — Creates rollups Define the output contract before moving to the next owner.

  7. 07

    Query engine — Scans ranges Confirm the result and emit the evidence needed to reconcile it.

04

Decision table

Make the trade-offs explicit.

DecisionDefensible positionCost to acknowledge
Primary mechanismRaw retention preserves flexibility; downsampling controls cost by discarding detail.The stronger guarantee usually adds coordination, latency, state, or operational work.
Sync vs. asyncKeep only correctness-critical confirmation synchronous. Move derived views, notifications, analytics, and cleanup behind a durable boundary.Async work needs idempotency, lag monitoring, replay, and a product definition for partial completion.
Simple vs. scaledBegin with one logical owner and a clear API. Partition or replicate only the resource proven to be the first bottleneck.Migration requires stable identities, versioned contracts, backfill, and a rollback path.
05

Failure review

Design the recovery path.

DetectBoundRetry safelyReconcileLearn

Topic-specific risk

Unbounded labels and hot series overwhelm indexes and partitions.

Response

Persist enough identity and state to distinguish retry, resume, compensation, and operator repair.

Dependency timeout

A timeout is ambiguous: the remote side may have failed, succeeded, or still be running.

Response

Use deadlines, bounded backoff with jitter, idempotency keys, and a status or reconciliation path.

Overload or skew

Average capacity can look healthy while a tenant, key, partition, region, or expensive request saturates one owner.

Response

Expose queue depth and hot-key share, apply backpressure, isolate tenants, and degrade optional work before correctness.

06

Evidence + level bar

Prove the design can be operated.

Core signals

Health of the promise

Measure user-visible latency or freshness, correctness drift, saturation, retry volume, and time to recover. Alert on the failed promise—not only CPU.

Mid-level

Complete and clear

Finish the happy path, identify the state owner, choose reasonable building blocks, and explain one scale mechanism.

Senior

Trade-offs and failure

Separate read and write paths, define consistency, explain partitioning, and make duplicate or partial failure safe.

Staff+

Evolution and operations

Discuss multi-region boundaries, migration, tenant isolation, capacity, observability, and how the architecture changes over time.

07

Interview language

Open the deep dive with a claim.

“For Time-series metrics ingestion, the decision I want to make explicit is this: Raw retention preserves flexibility; downsampling controls cost by discarding detail. I’ll trace the state-changing path first, show where the result becomes durable, then test the design against the highest-risk failure and our target scale.”

08 · Retrieval check

Can you defend it without the page?

  1. For Time-series metrics ingestion, where is the correctness boundary and which failure would you test first?
  2. Which component owns committed truth, and what event or response proves the commit?
  3. Where is the first scaling or coordination bottleneck under the stated envelope?
  4. What happens after an ambiguous timeout or duplicate operation?
  5. Which complexity would you remove at one hundredth of the scale?