FAQ
What is Cosmopilot?
Cosmopilot is a Kubernetes operator that deploys and manages Cosmos-SDK-based
blockchain nodes. It automates node deployment, upgrades, disk resizing, API exposure,
backups (with integrity validation) and more. See the Architecture
overview for how it works.
What chains are supported?
Any chain built on the Cosmos SDK that meets the required CLI commands and API endpoints. Check the Chain Compatibility page for the requirements and the list of tested chains.
What's the difference between a ChainNode and a ChainNodeSet?
A ChainNode manages a single node. A ChainNodeSet manages multiple nodes organized
into groups (for example full nodes, sentries, validators) with shared genesis,
services and ingresses. A ChainNodeSet creates and owns ChainNode resources behind
the scenes. See Deploy a Node and
Deploy a Node Set.
Do I need cert-manager?
Only if you keep admission webhooks enabled (the default) or expose endpoints with TLS.
You can install without cert-manager by setting webHooksEnabled=false, at the cost of
the validation webhooks. See Installation.
Is each node a StatefulSet?
No. Each ChainNode is backed by a single Pod plus its own PVC(s), Service and
configuration, which gives the operator fine-grained control over the node lifecycle.
How are nodes upgraded?
Cosmopilot can monitor governance for upgrade proposals and upgrade nodes
automatically when the upgrade height is reached, and also supports manually scheduled
upgrades. See Upgrades.
How does backup work?
Cosmopilot takes volume snapshots on a configurable schedule with retention
policies, can optionally verify their integrity by starting a throwaway node from the
snapshot, and can export snapshots as tar archives to GCS, Amazon S3, or
S3-compatible storage. See
Persistence & Backup.
Can I expose my node's API publicly?
Yes — RPC, LCD, gRPC and EVM endpoints can be exposed via Ingress or Gateway API, optionally fronted by CosmoGuard for access control and caching. See Exposing Endpoints.
How do I monitor nodes?
Each node exposes Prometheus metrics on port 26660. Cosmopilot does not create
ServiceMonitor resources for you — see Monitoring & Observability
for a ready-to-use example.
Can I run more than one operator instance?
Yes. You can shard work across multiple operator instances using workerName, and
control reconcile concurrency with workerCount. See
Configuration.
How do I uninstall Cosmopilot?
See the Uninstall guide. Note that removing the CRDs
deletes all ChainNode/ChainNodeSet resources, so plan accordingly.
Where do I report bugs or ask for help?
Open an issue at github.com/voluzi/cosmopilot/issues or email dev@voluzi.com.