docs-v2/content/influxdb/clustered/reference/internals/storage-engine.md

188 lines
7.9 KiB
Markdown

---
title: InfluxDB v3 storage engine architecture
description: >
The InfluxDB v3 storage engine is a real-time, columnar database optimized for
time series data that supports infinite tag cardinality, real-time queries,
and is optimized to reduce storage cost.
weight: 103
menu:
influxdb_clustered:
name: Storage engine architecture
parent: InfluxDB internals
influxdb/clustered/tags: [storage, internals]
related:
- /influxdb/clustered/admin/scale-cluster/
- /influxdb/clustered/admin/custom-partitions/
---
The InfluxDB v3 storage engine is a real-time, columnar database optimized for
time series data built in [Rust](https://www.rust-lang.org/) on top of
[Apache Arrow](https://arrow.apache.org/) and
[DataFusion](https://arrow.apache.org/datafusion/user-guide/introduction.html).
It supports infinite tag cardinality (number of unique tag values), real-time
queries, and is optimized to reduce storage cost.
- [Storage engine diagram](#storage-engine-diagram)
- [Storage engine components](#storage-engine-components)
- [Router](#router)
- [Ingester](#ingester)
- [Querier](#querier)
- [Catalog](#catalog)
- [Object store](#object-store)
- [Compactor](#compactor)
- [Garbage collector](#garbage-collector)
## Storage engine diagram
{{< svg "/static/svgs/v3-storage-architecture.svg" >}}
## Storage engine components
- [Router](#router)
- [Ingester](#ingester)
- [Querier](#querier)
- [Catalog](#catalog)
- [Object store](#object-store)
- [Compactor](#compactor)
- [Garbage collector](#garbage-collector)
### Router
The Router (also known as the Ingest Router) parses incoming line
protocol and then routes it to [Ingesters](#ingester).
To ensure write durability, the Router replicates data to two or more of the
available Ingesters.
##### Router scaling strategies
The Router can be scaled both [vertically](/influxdb/clustered/admin/scale-cluster/#vertical-scaling)
and [horizontally](/influxdb/clustered/admin/scale-cluster/#horizontal-scaling).
Horizontal scaling increases write throughput and is typically the most
effective scaling strategy for the Router.
Vertical scaling (specifically increased CPU) improves the Router's ability to
parse incoming line protocol with lower latency.
### Ingester
The Ingester processes line protocol submitted in write requests and persists
time series data to the [Object store](#object-store).
In this process, the Ingester does the following:
- Queries the [Catalog](#catalog) to identify where data should be persisted and
to ensure the schema of the line protocol is compatible with the
[schema](/influxdb/clustered/reference/glossary/#schema) of persisted data.
- Accepts or [rejects](/influxdb/clustered/write-data/troubleshoot/#troubleshoot-rejected-points)
points in the write request and generates a [response](/influxdb/clustered/write-data/troubleshoot/).
- Processes line protocol and persists time series data to the
[Object store](#object-store) in Apache Parquet format. Each Parquet file
represents a _partition_--a logical grouping of data.
- Makes [yet-to-be-persisted](/influxdb/clustered/reference/internals/durability/#data-ingest)
data available to [Queriers](#querier) to ensure leading edge data is included
in query results.
- Maintains a short-term [write-ahead log (WAL)](/influxdb/clustered/reference/internals/durability/)
to prevent data loss in case of a service interruption.
##### Ingester scaling strategies
The Ingester can be scaled both [vertically](/influxdb/clustered/admin/scale-cluster/#vertical-scaling)
and [horizontally](/influxdb/clustered/admin/scale-cluster/#horizontal-scaling).
Vertical scaling increases write throughput and is typically the most
effective scaling strategy for the Ingester.
### Querier
The Querier handles query requests and returns query results for requests.
It supports both SQL and InfluxQL through
[Apache Arrow DataFusion](https://arrow.apache.org/datafusion/user-guide/introduction.html).
#### Query life cycle
At query time, the querier:
1. Receives the query request and builds a query plan.
2. Queries the [Ingesters](#ingester) to:
- ensure the schema assumed by the query plan matches the schema of written data
- include recently written, [yet-to-be-persisted](/influxdb/clustered/reference/internals/durability/#data-ingest)
data in query results
3. Queries the [Catalog](#catalog) to find partitions in the [Object store](#object-store)
that contain the queried data.
4. Reads partition Parquet files that contain the queried data and scans each
row to filter data that matches predicates in the query plan.
5. Performs any additional operations (for example: deduplicating, merging, and sorting)
specified in the query plan.
6. Returns the query result to the client.
##### Querier scaling strategies
The Querier can be scaled both [vertically](/influxdb/clustered/admin/scale-cluster/#vertical-scaling)
and [horizontally](/influxdb/clustered/admin/scale-cluster/#horizontal-scaling).
Horizontal scaling increases query throughput to handle more concurrent queries.
Vertical scaling improves the Querier's ability to process computationally
intensive queries.
### Catalog
The Catalog is a PostgreSQL-compatible relational database that stores metadata
related to your time series data including schema information and physical
locations of partitions in the [Object store](#object-store).
It fulfills the following roles:
- Provides information about the schema of written data.
- Tells the [Ingester](#ingester) what partitions to persist data to.
- Tells the [Querier](#querier) what partitions contain the queried data.
##### Catalog scaling strategies
Scaling strategies available for the Catalog depend on the PostgreSQL-compatible
database used to run the catalog. All support
[vertical scaling](/influxdb/clustered/admin/scale-cluster/#vertical-scaling).
Most support [horizontal scaling](/influxdb/clustered/admin/scale-cluster/#horizontal-scaling)
for redundancy and failover.
### Object store
The Object store contains time series data in [Apache Parquet](https://parquet.apache.org/) format.
Each Parquet file represents a partition.
By default, InfluxDB partitions tables by day, but you can
[customize the partitioning strategy](/influxdb/clustered/admin/custom-partitions/).
Data in each Parquet file is sorted, encoded, and compressed.
##### Object store scaling strategies
Scaling strategies available for the Object store depend on the underlying
object storage services used to run the object store.
Most support [horizontal scaling](/influxdb/clustered/admin/scale-cluster/#horizontal-scaling)
for redundancy, failover, and increased capacity.
### Compactor
The Compactor processes and compresses partitions in the [Object store](#object-store)
to continually optimize storage.
It then updates the [Catalog](#catalog) with locations of compacted data.
##### Compactor scaling strategies
The Compactor can be scaled both [vertically](/influxdb/clustered/admin/scale-cluster/#vertical-scaling)
and [horizontally](/influxdb/clustered/admin/scale-cluster/#horizontal-scaling).
Because compaction is a compute-heavy process, vertical scaling (especially
increasing the available CPU) is the most effective scaling strategy for the Compactor.
Horizontal scaling increases compaction throughput, but not as efficiently as
vertical scaling.
### Garbage collector
The Garbage collector runs background jobs that evict expired or deleted data,
remove obsolete compaction files, and reclaim space in both the [Catalog](#catalog) and the
[Object store](#object-store).
##### Garbage collector scaling strategies
The Garbage collector is not designed for distributed load and should _not_ be
scaled horizontally. The Garbage collector does not perform CPU- or
memory-intensive work, so [vertical scaling](/influxdb/clustered/admin/scale-cluster/#vertical-scaling)
should only be considered only if you observe very high CPU usage or
if the container regularly runs out of memory.