Scale your InfluxDB cluster

InfluxDB Clustered lets you scale individual components of your cluster both vertically and horizontally to match your specific workload. Use the AppInstance resource defined in your influxdb.yml to manage resources available to each component.

• Scaling strategies
  • Vertical scaling
  • Horizontal scaling
• Scale your cluster as a whole
• Scale components in your cluster
  • Horizontally scale a component
  • Vertically scale a component
  • Apply your changes
• Recommended scaling strategies per component

Scaling strategies

The following scaling strategies can be applied to components in your InfluxDB cluster.

Vertical scaling

Vertical scaling (also known as “scaling up”) involves increasing the resources (such as RAM or CPU) available to a process or system. Vertical scaling is typically used to handle resource-intensive tasks that require more processing power.

Horizontal scaling

Horizontal scaling (also known as “scaling out”) involves increasing the number of nodes or processes available to perform a given task. Horizontal scaling is typically used to increase the amount of workload or throughput a system can manage, but also provides additional redundancy and failover.

Scale your cluster as a whole

Scaling your entire InfluxDB Cluster is done by scaling your Kubernetes cluster and is managed outside of InfluxDB. The process of scaling your entire Kubernetes cluster depends on your underlying Kubernetes provider. You can also use Kubernetes autoscaling to automatically scale your cluster as needed.

Scale components in your cluster

The following components of your InfluxDB cluster are scaled by modifying properties in your AppInstance resource:

• Ingester
• Querier
• Compactor
• Router
• Garbage collector
• Catalog service

apiVersion: kubecfg.dev/v1alpha1
kind: AppInstance
# ...
spec:
  package:
    spec:
      # The following settings tune the various pods for their cpu/memory/replicas
      # based on workload needs. Only uncomment the specific resources you want
      # to change. Anything left commented will use the package default.
      resources:
        ingester:
          requests:
            cpu: 
INGESTER_CPU_MIN
            memory: 
INGESTER_MEMORY_MIN
            replicas: 
INGESTER_REPLICAS # Default is 3
          limits:
            cpu: 
INGESTER_CPU_MAX
            memory: 
INGESTER_MEMORY_MAX
        compactor:
          requests:
            cpu: 
COMPACTOR_CPU_MIN
            memory: 
COMPACTOR_MEMORY_MIN
            replicas: 
COMPACTOR_REPLICAS # Default is 1
          limits:
            cpu: 
COMPACTOR_CPU_MAX
            memory: 
COMPACTOR_MEMORY_MAX
        querier:
          requests:
            cpu: 
QUERIER_CPU_MIN
            memory: 
QUERIER_MEMORY_MIN
            replicas: 
QUERIER_REPLICAS # Default is 1          
          limits:
            cpu: 
QUERIER_CPU_MAX
            memory: 
QUERIER_MEMORY_MAX
        router:
          requests:
            cpu: 
ROUTER_CPU_MIN
            memory: 
ROUTER_MEMORY_MIN
            replicas: 
ROUTER_REPLICAS # Default is 1
          limits:
            cpu: 
ROUTER_CPU_MAX
            memory: 
ROUTER_MEMORY_MAX
        garbage-collector:
          requests:
            cpu: 
GC_CPU_MIN
            memory: 
GC_MEMORY_MIN
          limits:
            cpu: 
GC_CPU_MAX
            memory: 
GC_MEMORY_MAX

# ...
  resources:
    ingester:
      requests:
        cpu: 
INGESTER_CPU_MIN
        memory: 
INGESTER_MEMORY_MIN
        replicas: 
INGESTER_REPLICAS # Default is 3
      limits:
        cpu: 
INGESTER_CPU_MAX
        memory: 
INGESTER_MEMORY_MAX
    compactor:
      requests:
        cpu: 
COMPACTOR_CPU_MIN
        memory: 
COMPACTOR_MEMORY_MIN
        replicas: 
COMPACTOR_REPLICAS # Default is 1
      limits:
        cpu: 
COMPACTOR_CPU_MAX
        memory: 
COMPACTOR_MEMORY_MAX
    querier:
      requests:
        cpu: 
QUERIER_CPU_MIN
        memory: 
QUERIER_MEMORY_MIN
        replicas: 
QUERIER_REPLICAS # Default is 1          
      limits:
        cpu: 
QUERIER_CPU_MAX
        memory: 
QUERIER_MEMORY_MAX
    router:
      requests:
        cpu: 
ROUTER_CPU_MIN
        memory: 
ROUTER_MEMORY_MIN
        replicas: 
ROUTER_REPLICAS # Default is 1
      limits:
        cpu: 
ROUTER_CPU_MAX
        memory: 
ROUTER_MEMORY_MAX
    garbage-collector:
      requests:
        cpu: 
GC_CPU_MIN
        memory: 
GC_MEMORY_MIN
      limits:
        cpu: 
GC_CPU_MAX
        memory: 
GC_MEMORY_MAX

Related Kubernetes documentation

• CPU resource units
• Memory resource units

Horizontally scale a component

To horizontally scale a component in your InfluxDB cluster, increase or decrease the number of replicas for the component and apply the change.

For example–to horizontally scale your Ingester:

apiVersion: kubecfg.dev/v1alpha1
kind: AppInstance
# ...
spec:
  package:
    spec:
      resources:
        ingester:
          requests:
            # ...
            replicas: 6

# ...
resources:
  ingester:
    requests:
      # ...
      replicas: 6

Vertically scale a component

To vertically scale a component in your InfluxDB cluster, increase or decrease the CPU and memory resource units to assign to component pods and apply the change.

apiVersion: kubecfg.dev/v1alpha1
kind: AppInstance
# ...
spec:
  package:
    spec:
      resources:
        ingester:
          requests:
            cpu: "500m"
            memory: "512MiB"
            # ...
          limits:
            cpu: "1000m"
            memory: "1024MiB"

# ...
resources:
  ingester:
    requests:
      cpu: "500m"
      memory: "512MiB"
      # ...
    limits:
      cpu: "1000m"
      memory: "1024MiB"

Apply your changes

After modifying the AppInstance resource, use kubectl apply to apply the configuration changes to your cluster and scale the updated components.

kubectl apply \
  --filename myinfluxdb.yml \
  --namespace influxdb

helm upgrade \
  influxdata/influxdb3-clustered \
  -f ./values.yml \
  --namespace influxdb

Recommended scaling strategies per component

• Router
• Ingester
• Querier
• Compactor
• Garbage collector
• Catalog store
• Catalog service
• Object store

Router

The Router can be scaled both vertically and horizontally.

• Recommended: 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.

Router latency

Latency of the Router’s write endpoint is directly impacted by:

• Ingester latency–the router calls the Ingester during a client write request
• Catalog latency during schema validation

Ingester

The Ingester can be scaled both vertically and horizontally.

• Recommended: Vertical scaling is typically the most effective scaling strategy for the Ingester. Compared to horizontal scaling, vertical scaling not only increases write throughput but also lessens query, catalog, and compaction overheads as well as Object store costs.
• Horizontal scaling can help distribute write load but comes with additional coordination overhead.

Ingester storage volume

Ingesters use an attached storage volume to store the Write-Ahead Log (WAL). With more storage available, Ingesters can keep bigger WAL buffers, which improves query performance and reduces pressure on the Compactor. Storage speed also helps with query performance.

Configure the storage volume attached to Ingester pods in the spec.package.spec.ingesterStorage property of your AppInstance resource or, if using Helm, the ingesterStorage property of your values.yaml.

apiVersion: kubecfg.dev/v1alpha1
kind: AppInstance
# ...
spec:
  package:
    spec:
      # ...
      ingesterStorage:
        # (Optional) Set the storage class. This will differ based on the K8s
        #environment and desired storage characteristics.
        # If not set, the default storage class is used.
        storageClassName: 
STORAGE_CLASS
        # Set the storage size (minimum 2Gi recommended)
        storage: 
STORAGE_SIZE

# ...
ingesterStorage:
  # (Optional) Set the storage class. This will differ based on the K8s
  #environment and desired storage characteristics.
  # If not set, the default storage class is used.
  storageClassName: 
STORAGE_CLASS
  # Set the storage size (minimum 2Gi recommended)
  storage: 
STORAGE_SIZE

Querier

The Querier can be scaled both vertically and horizontally.

• Recommended: Vertical scaling improves the Querier’s ability to process concurrent or computationally intensive queries, and increases the effective cache capacity.
• Horizontal scaling increases query throughput to handle more concurrent queries. Consider horizontal scaling if vertical scaling doesn’t adequately address concurrency demands or reaches the hardware limits of your underlying nodes.

Compactor

• Recommended: Maintain 1 Compactor pod and use vertical scaling (especially increasing the available CPU) for the Compactor.
• Because compaction is a compute-heavy process, horizontal scaling increases compaction throughput, but not as efficiently as vertical scaling.

Garbage collector

The Garbage collector is a lightweight process that typically doesn’t require significant system resources.

• Don’t horizontally scale the Garbage collector; it isn’t designed for distributed load.
• Consider vertical scaling only if you observe consistently high CPU usage or if the container regularly runs out of memory.

Catalog store

The Catalog store is a PostgreSQL-compatible database that stores critical metadata for your InfluxDB cluster. An underprovisioned Catalog store can cause write outages and system-wide performance issues.

• Scaling strategies depend on your specific PostgreSQL implementation
• All PostgreSQL implementations support vertical scaling
• Most implementations support horizontal scaling for improved redundancy and failover

Catalog service

The Catalog service (iox-shared-catalog statefulset) caches and manages access to the Catalog store.

• Recommended: Maintain exactly 3 replicas of the Catalog service for optimal redundancy. Additional replicas are discouraged.
• If performance improvements are needed, use vertical scaling.

Object store

The Object store contains time series data in Parquet format.

Scaling strategies depend on the underlying object storage services used. Most services support horizontal scaling for redundancy, failover, and increased capacity.