View and apply resource recommendations, investigate container health issues, and query Kubernetes events from the CLI.

Managing Optimizations

Once the Write Operator is collecting data, the DevZero platform generates resource optimization recommendations. The CLI lets you review, apply, and track these — alongside tools for investigating container health issues and Kubernetes events.

Recommendations

View and apply workload resource recommendations. Alias: dz recommendations.

# List all recommendations
dz recs list

# Only pending or applied
dz recs pending
dz recs applied

# Details for one workload
dz recs get my-api --namespace production

Flag	Short	Default	Description
`--cluster`	`-c`		Cluster ID (interactive selector if omitted)
`--namespace`	`-n`		Filter by namespace
`--since`		`7d`	Lookback window (`30d`, `4w`, `168h`)
`--status`		`all`	Filter: `applied`, `pending`, `all` (on `list`)

dz recs applied also shows aggregate monthly cost savings across all applied recommendations.

Apply Workflow

Review

dz recs get my-api --namespace production

Shows per-container before/after requests and limits, the reasoning behind the recommendation, and estimated monthly savings.

Apply

dz recs apply my-api --namespace production

The CLI shows the before/after diff and asks for confirmation. Pass --yes to skip the prompt.

Verify

dz recs applied --namespace production

dz recs apply queues the recommendation for the Write Operator (dakr) running in-cluster. Changes are applied during its next reconciliation loop, typically within seconds.

Apply Flags

Flag	Short	Description
`--kind`	`-k`	Workload kind (e.g. `Deployment`, `StatefulSet`)
`--yes`	`-y`	Skip confirmation prompt

Events

Query recent Kubernetes events for a connected cluster. Events are grouped by reason, workload, and namespace by default.

# Grouped events from the last 24h (default)
dz events

# Warning events from the last hour
dz events --warnings --since 1h

# Filter by reason
dz events --reason OOMKilling --namespace production

# Top-N aggregation summary
dz events --summary

# Individual events (ungrouped)
dz events --raw

Flag	Short	Default	Description
`--cluster`	`-c`		Cluster ID
`--namespace`	`-n`		Filter by namespace (repeatable)
`--reason`			Filter by reason, e.g. `OOMKilling`, `FailedScheduling` (repeatable)
`--warnings`	`-w`	`false`	Warnings only
`--since`		`24h`	Lookback window
`--raw`		`false`	Individual events (mutually exclusive with `--summary`)
`--summary`		`false`	Top-N aggregation
`--limit`		`50`	Max results

Container Health

Surface OOM kills, crash loops, and CPU throttling across a cluster or for a specific workload.

# Cluster-wide summary
dz health

# Sort by crash loops instead of OOMs
dz health --sort crashloop

# Detailed health for a specific workload
dz health my-api --namespace production

Without arguments, dz health shows a cluster-wide table of workloads with health issues. With a workload name, it shows a detailed breakdown — last OOM timestamp, memory at time of kill, throttled period counts, and crash loop history.

Flag	Short	Default	Description
`--cluster`	`-c`		Cluster ID
`--namespace`	`-n`		Namespace (for workload detail view)
`--sort`		`oom`	Sort: `oom`, `crashloop`, `throttle`, `total`

Managing Optimizations

On this page