Remote Operator Debug

Run kubectl and other commands on customer collectors via remote-operator.

Prerequisites

•aws-sso-login skill: AWS authentication
•remote-operator binary in PATH

Environment URLs

Environment	Remote Operator Address
prod	`https://remote-operator-manager.ue1.prod.plat.sm-svc.com`
staging	`https://remote-operator-manager.ue1.staging.plat.sm-svc.com`

Workflow

Step 1: Get Org ID and Environment

Ask user for:

•Clerk org ID (e.g., org_345YDItRYVf46uffseTpSKBXNF1)
•Environment (prod/staging)

Step 2: List Deployments

bash

remote-operator -a <ro_address> -o <org_id> manage ls

Output shows:

•Deployment names
•Instance names (needed for commands)
•Pod status

Step 3: Run Commands

Basic pattern:

bash

remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- <command>

Common Commands

Get Logs

bash

# Main collector
remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl logs -n sawmills deployment/sawmills-collector -c main-collector --tail 200

# HAProxy
remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl logs -n sawmills deployment/sawmills-collector -c haproxy --tail 100

# Telemetry collector
remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl logs -n sawmills deployment/sawmills-collector -c telemetry-collector --tail 100

Check Pod Status

bash

remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl get pods -n sawmills

Describe Deployment

bash

remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl describe deployment sawmills-collector -n sawmills

Get Helm Values

bash

remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- helm get values sawmills-collector -n sawmills

Restart Pods

bash

remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl rollout restart deployment/sawmills-collector -n sawmills

Check Events

bash

remote-operator -a <ro_address> -o <org_id> manage run \
  -d <deployment> --instance-name <instance> \
  -- kubectl get events -n sawmills --sort-by='.lastTimestamp'

Filtering Logs

bash

# Grep for errors
remote-operator ... -- kubectl logs ... | grep -i error

# Filter by time (last 5 minutes)
remote-operator ... -- kubectl logs ... --since=5m

# Follow logs (streaming)
remote-operator ... -- kubectl logs ... -f

Namespace

Default namespace is sawmills. Service account has limited permissions:

•Can access sawmills namespace
•Cannot list pods across all namespaces (-A flag fails)

Troubleshooting

Issue	Solution
`executable file not found`	Command not available in container
`Forbidden`	Service account lacks permission
`No resources found`	Wrong namespace or no matching pods
`timeout`	Network issues or large output

Example Session

bash

# 1. Set variables
ORG=org_345YDItRYVf46uffseTpSKBXNF1
RO=https://remote-operator-manager.ue1.prod.plat.sm-svc.com

# 2. List deployments
remote-operator -a $RO -o $ORG manage ls

# 3. Get logs from specific instance
remote-operator -a $RO -o $ORG manage run \
  -d SawmillsPoc --instance-name 7fc25w8 \
  -- kubectl logs -n sawmills deployment/sawmills-collector -c main-collector --tail 100

Notes

•Instance name changes on redeployment; re-run manage ls to get current
•Large log output may be truncated; use --tail to limit
•Some commands like logs may default to wrong container; specify -c