InfluxDB

InfluxDB time-series platforms, built and run

InfluxDB is built for data that accumulates over time: server, equipment, IoT and application metrics. We design everything from edition choice (2.x / 3 Core / 3 Enterprise) to Telegraf collection, retention and downsampling, backups and Grafana visualisation.

InfluxDB consultation Compare versions

Pipeline

From collection to visualisation

The classic TIG (Telegraf · InfluxDB · Grafana) pipeline.

Telegraf agents collect server, database, network and equipment metrics and send them to InfluxDB; Grafana provides dashboards and alerts. Data that must be kept long-term is downsampled into a separate bucket or database to cut storage cost and query time.

Control cardinality through measurement, tag and field design
Separate retention for raw and aggregated data
Tune write load via collection interval and batch size
Least-privilege tokens and TLS

Sourcesservers · DBs · IoT · apps

collect

Telegrafinput → processor → output

write

InfluxDBraw (short retention)

Downsampledaggregates (long retention)

query

Grafanadashboards · alerts

Backup (object store / CLI)Tokens · TLSCapacity monitoring

Figure. Telegraf–InfluxDB–Grafana reference pipeline

Versions

Choosing a version (as of October 2026)

Same name, very different products: 2.x and 3.x differ in storage engine, query languages and operations.

InfluxDB editions compared
Item	InfluxDB 2.x OSS	InfluxDB 3 Core	InfluxDB 3 Enterprise
Licence	Open source	Open source	Commercial
Query	Flux, InfluxQL	SQL, InfluxQL	SQL, InfluxQL
Default port · CLI	8086 · `influx`	8181 · `influxdb3`	8181 · `influxdb3`
Data unit	Organization / Bucket	Database / Table	Database / Table
Retention	Per bucket, changeable	Set at database creation, immutable afterwards	Changeable; per-table retention
Backup	`influx backup / restore`	No built-in command — copy object-store files in a defined order	Object-store based
High availability	Single node	Single node	HA, read replicas and other cluster features

Source: docs.influxdata.com (checked October 2026). The Docker influxdb:latest tag now points to 3 Core, so pin a version tag if you run 2.x. Versions and features change often; we re-check the official docs at adoption time.

Scope

InfluxDB service scope

Install & design

Edition choice, package/Docker installs, storage (local/object) design, schema and cardinality design.

Telegraf collection

Input plugins, outputs.influxdb_v2 / outputs.influxdb_v3, buffer and batch tuning.

Retention & downsampling

Raw vs aggregate retention; aggregation with 2.x Tasks (Flux) or 3.x processing-engine plugins.

Backup & recovery

2.x CLI backups; 3.x object-store copy procedure (stop service, restore in reverse order) and testing.

High availability

3 Enterprise clustering, or dual-write/replication alternatives for OSS.

Tuning

Write batching and compression, query time ranges and tag filters, memory/cache settings, high-cardinality fixes.

Grafana

Data-source setup (SQL recommended for 3.x), dashboards, alert rules, operations panels.

Migration

1.x/2.x → 3.x migration plans, Flux → SQL/InfluxQL query conversion, parallel running.

Examples

Configuration examples

InfluxDB 3 Core — create a database (30-day retention)

# Retention can only be set at creation time (Core)
influxdb3 create database --retention-period 30d \
  --token "$INFLUXDB3_AUTH_TOKEN" sensors

# Create with no expiry
influxdb3 create database --token "$INFLUXDB3_AUTH_TOKEN" sensors_raw

Telegraf → InfluxDB 3 output (telegraf.conf)

# Telegraf 1.38+: dedicated InfluxDB 3 output plugin
[[outputs.influxdb_v3]]
  urls     = ["https://influx.example.internal:8181"]
  token    = "${INFLUX_TOKEN}"
  database = "sensors"

[[inputs.cpu]]
  percpu = false
  totalcpu = true

Because 3 Core retention cannot be changed later (it requires a new database and a data migration), it is safest to design separate raw and aggregate databases from the start.

FAQ

Frequently asked questions

Starting fresh — 2.x or 3.x?

Choose 3.x for SQL analytics and object storage; 2.x is pragmatic if you want to keep existing Flux assets and proven 2.x operations. We decide together based on retention-change needs, HA requirements (Enterprise or not) and existing dashboards.

Does InfluxDB 3 Core have a backup command?

As of October 2026, 3 Core has no built-in backup command; you back up by copying files from object storage (or the local data path) in the order the official docs describe, and restore in reverse order with the service stopped.

How do we connect Grafana to 3.x?

In Grafana’s InfluxDB data source, choose SQL (or InfluxQL) as the query language. Flux is not used with 3.x.

What does “high cardinality” mean?

Too many unique tag-value combinations (series). Using ever-growing values such as user or request IDs as tags can cause performance and memory problems, so we adjust the schema, e.g. by moving them to fields.

Shall we start by reviewing your cloud and databases as they are today?

Tell us your clouds, database versions, topology and concerns, and we will propose an assessment scope and approach.

Get in touch khy061473@gmail.com