Skip to main content
falkordb-admin drives the same Admin Server API as the console, which makes it a good fit for scripting, CI, and support sessions.

Run it

The CLI ships inside the Admin Server and Admin UI images, so the quickest way to use it is from a pod that is already running:
You can also build the standalone CLI image from a checkout of the repository:

Connect

The CLI keeps the server URL and session cookie in ~/.falkordb/admin-cli/config.json.
Use --password-stdin and pipe the password in, so it never lands in your shell history or in a process listing.
Add --output json to any command to get machine-readable output.
Commands that act on a deployment require --namespace. Only the top-level listings, such as deployments list, treat it as optional.

Common tasks

Run falkordb-admin --help for the full command list.

Change a deployment

The CLI changes deployments with the deployment body:
  • deployments create --file takes a body with schemaVersion: 1.
  • deployments update --file takes a merge patch with schemaVersion: 1; deployments replace --file takes a whole body. Add --generation with the metadata.generation you read, so a change made since is refused instead of overwritten, and --dry-run to see the plan without applying it.
  • deployments scale, deployments upgrade, and operations start, stop, vertical-scale, volume-expand and reconfigure read the deployment and send only what changes, with its generation. scale takes --replicas (nodes, per shard when sharded), --shards, and --component falkordb-sent for the sentinels; as before, sentinels keep at least 3 replicas and a deployment with sentinels or shards keeps at least 2 nodes. --generation makes them refuse the change if the deployment moved past the generation you name.
  • vertical-scale --cpu and --memory set both the request and the limit; --requests-cpu, --limits-memory and the like set one side.
  • reconfigure takes --parameter key=value once per setting, so values may contain commas. --parameter key on its own sets a falkordbe.* setting back to its default, and key= empties one whose default is empty, or a redis.conf directive the pods apply without a restart. See how settings are applied.
  • Reconfiguring Sentinel (--component falkordb-sent), restart, switchover, rebalance and reshard still run as KubeBlocks operations.
  • reshard moves the slots in --slots (such as 0-999,5000) to the shard in --to, and with --from only if that shard owns them.
  • rebalance spreads the hash slots evenly across the shards of a sharded deployment. Adding or removing shards already moves slots; use it when the spread is uneven afterwards.
A file without schemaVersion still works with create and update. The Admin Server marks that request deprecated, and the CLI prints a warning to stderr.

Call an endpoint directly

Anything the CLI does not wrap yet is reachable through raw, which reuses the stored session:

Next steps

falkordb-admin --help lists every command, and falkordb-admin <command> --help every flag for that command.

Admin Server API

The REST API the CLI talks to

Installer reference

Options for install.sh and uninstall.sh