Node Configurations

This page provides a comprehensive guide to configuring ChainNode (or ChainNodeSet group) resources in Cosmopilot, covering various options such as resource management, affinity, node configurations, and more. Some of below configuration fields are just reflected in corresponding TOML config files, and were created to simplify configuration, but its always possible to override TOML config files as explained below for more advanced configurations.

State-Sync Snapshots

config:
  stateSync:
    snapshotInterval: 250
    snapshotKeepRecent: 5 # optional. Defaults to 2.

Overriding TOML Config Files

When a ChainNode is initialized for the first time, Cosmopilot generates all necessary configuration files (stored in the config directory) and saves them in a ConfigMap with the same name as the ChainNode. Among these files, Cosmopilot modifies only app.toml and config.toml to apply most settings. However, you can override these configurations — or settings in additional files — using the .spec.config.override field.

Although these configuration files are typically in TOML format, they should be defined in YAML format when using the override field. Cosmopilot will automatically convert the YAML into valid TOML syntax before applying the changes.

Only the specific values provided in the override field are changed. All other settings remain as their default values, as defined by the application. If you need to restore a configuration to its default state, simply remove it from the override field.

config:
  override:
    app.toml:
      app-db-backend: pebbledb
      iavl-lazy-loading: true
      min-retain-blocks: 500000
      minimum-gas-prices: 0.025unibi
      pruning: custom
      pruning-interval: "10"
      pruning-keep-recent: "100"
    config.toml:
      db_backend: pebbledb

Setting Pod Resources

Configure resource requests and limits for the main application container using the resources field:

resources:
  requests:
    cpu: "500m"
    memory: "1Gi"
  limits:
    cpu: "1"
    memory: "2Gi"

Setting Image Pull Secrets

To use private container images, you can specify image pull secrets:

config:
  imagePullSecrets:
  - name: my-private-registry-secret

Node Selector and Affinity

You can control where the pod runs using node selectors and affinity rules:

Node Selector

nodeSelector:
  disktype: ssd

Affinity

affinity:
  podAntiAffinity:
    requiredDuringSchedulingIgnoredDuringExecution:
    - labelSelector:
        matchLabels:
        chain-id: nibiru-testnet-1
        group: fullnode
        nodeset: nibiru-testnet-1
      topologyKey: topology.gke.io/zone

Configuring Block Threshold

Cosmopilot provides a special feature to monitor block timestamps instead of relying solely on the catching_up field. This ensures more reliable health checks. By default, the block threshold is 15s. To configure it:

config:
  blockThreshold: 30s

Seed Mode

To enable seed mode, configure the following:

config:
  seedMode: true

Additional ENV Variables

You can pass additional environment variables to the main application container:

env:
  - name: CUSTOM_VAR_1
    value: custom_value_1
  - name: CUSTOM_VAR_2
    value: custom_value_2

Configuring Node Startup Time

The startup time corresponds to the startup probe timeout. It defaults to 1h. If the node does not get helthy within this period it will be restarted. In some cases, like when starting a node with huge data, this might not be enough. You can adjust adjust it, using the following:

config:
  startupTime: 3h

node-utils Resources

You can configure resource requests and limits for the node-utils container:

config:
  nodeUtilsResources:
    requests:
      cpu: "300m"
      memory: "100Mi"
    limits:
      cpu: "300m"
      memory: "100Mi"

The example above actually represents the defaults values.

Persisting Address Book File

By default, the address book file is not persisted accross restarts, and is rebuilt on every new start. To persist the node's address book file, enable the following option:

config:
  persistAddressBook: true

Enable EVM

If the blockchain network supports EVM, enable it with the following configuration:

evmEnabled: true

This will ensure that EVM RPC ports will be added to the node's service and will be available when exposing the endpoints.

Startup Flags

In some cases you might need to append additional startup flags to the main application. For example in osmosis nodes, the startup command will override some settings on both config.toml and app.toml, which will not work with Cosmopilot as it does manage those files. So an extra flag needs to be added to the main application, using the following:

config:
  runFlags: ["--reject-config-defaults=true"]

Additional Volumes

Cosmopilot creates a single main data PVC which will be mounted at /home/app/data. However, some applications might need to persist more data outside the data diretory. When possible, it is advisable to Override TOML config files to store additional data in the volume (in /home/app/data). However, this is not always possible. In those cases, additional volumes can be created to store that data as follows:

config:
  volumes:
  - name: wasm
    size: 1Gi
    path: /home/app/wasm
    deleteWithNode: true
  - name: ibc-08-wasm
    size: 1Gi
    path: /home/app/ibc_08-wasm
    deleteWithNode: true

Sidecar containers

In some networks, additional tools might need to run together with the node and have access to its data volume. On an environment where you have the possibility of having ReadWriteMany volumes this is not a problem, but on most cloud providers that is mostly not available, or extremely expensive. For this reason, Cosmopilot provides you with a way of appending additional containers to the same Pod as the node.

Please refer to SidecarSpec for full specification on all available fields.

Halt Height

In certain scenarios, you may need to stop a node at a specific block height. Cosmos-based nodes provide a halt-height configuration field in the app.toml file for this purpose. However, directly setting this field through TOML config overrides may not work as expected, as Cosmopilot will continuously attempt to restart the pod.

To properly handle this scenario, Cosmopilot offers its own .spec.config.haltHeight field, allowing you to set the desired halt height like this:

config:
  haltHeight: 12345

When configured this way, Cosmopilot will:

• Automatically set the halt-height in the app.toml file.
• Prevent the pod from restarting after the node stops at the specified height.

This ensures a smooth and controlled node shutdown without unintended restarts.