← Back to the chapter · Next detail →
Technical detail · Time-Series Platform Selection
Compare candidate release evidence
| Pattern | Use when | Be careful with | Release evidence |
|---|---|---|---|
| InfluxDB-style | The data is mostly timestamped measurements, dashboards filter by tags, and the team wants time-series ingestion and retention features. | Version-specific query paths, tag and field rules, wide or sparse schemas, and schema changes after first writes. | Line protocol contract, tag and field inventory, representative queries, retention check, backup or export plan. |
| TimescaleDB-style | The team needs SQL, joins to asset or customer metadata, PostgreSQL tools, continuous aggregates, and database governance. | Chunk sizing, indexes, aggregate refresh windows, PostgreSQL tuning, and raw retention interactions. | Hypertable DDL, indexes, query plans, continuous aggregate policy, retention policy, restore test. |
| Prometheus-style | The goal is infrastructure and application monitoring, scrape targets, metric labels, recording rules, and alerts. | High-cardinality labels, treating monitoring samples as business records, and assuming local storage is clustered. | Metric naming review, label budget, scrape coverage, alert rule tests, retention or remote-write decision. |
| QuestDB-style | The workload needs fast SQL analytics over ordered timestamp partitions and the primary filter is a time interval. | Designated timestamp choice, partition interval, late data, schema evolution, and integration needs. | Timestamp designation, partition policy, interval-scan proof, ingest replay, TTL or retention behavior. |