Skip to main content
Parallel Mode
This feature is currently in Beta. Update to the latest Hoop version to access it.

​
What You’ll Accomplish

Parallel Mode lets you run the same command across multiple connections at once. Instead of connecting to each server individually, you can:
  • Restart services across multiple servers simultaneously
  • Run the same query on multiple databases
  • Execute maintenance scripts on all production hosts
  • Check status across your entire infrastructure

​
How It Works

1

Select Connections

Choose multiple connections to run your command on
2

Enter Command

Type the command or query once
3

Execute

Hoop runs the command on all selected connections in parallel
4

View Results

See results from all connections in a unified view

​
Quick Start

​
Step 1: Open Parallel Mode

  1. Go to the Web Terminal in the Hoop Web App
  2. Click the Parallel Mode toggle in the top-right corner

​
Step 2: Select Connections

  1. Use the connection picker to select multiple connections
  2. Filter by type, tags, or name to find connections quickly
  3. Click connections to add them to your selection

​
Step 3: Run a Command

  1. Type your command in the input field
  2. Click Execute or press Enter
  3. Watch results stream in from all connections

​
Step 4: Review Results

The Execution Summary shows:
  • Success - Command completed successfully
  • Error - Command failed (click to see details)
  • Approval Required - Waiting for access request approval

​
Use Cases

​
1. Service Restarts

Restart a service across all application servers: 
sudo systemctl restart myapp
Select all app-server-* connections and execute once.

​
2. Database Health Checks

Check replication status on all database replicas: 
SELECT pg_is_in_recovery(), pg_last_wal_receive_lsn();
Run on db-replica-1, db-replica-2, db-replica-3 simultaneously.

​
3. Configuration Updates

Update a configuration file on multiple hosts: 
sed -i 's/old_value/new_value/g' /etc/myapp/config.yaml

​
4. Log Collection

Check recent errors across all servers: 
tail -100 /var/log/myapp/error.log | grep ERROR

​
5. Kubernetes Operations

Scale deployments across multiple clusters: 
kubectl scale deployment myapp --replicas=5
Run on connections for each Kubernetes cluster.

​
Working with Results

​
Execution Summary

The summary panel shows the status of each connection:
StatusIconMeaning
SuccessGreen checkCommand completed successfully
ErrorRed XCommand failed
RunningSpinnerCommand in progress
Pending ApprovalClockWaiting for access request

​
Viewing Output

Click on any connection in the summary to see its full output:
  • stdout - Standard output from the command
  • stderr - Error output (if any)
  • Exit code - Command exit status

​
Comparing Results

Results are displayed side-by-side, making it easy to:
  • Spot differences between servers
  • Identify which servers have issues
  • Verify consistent state across infrastructure

​
Filtering and Selection

​
Filter by Type

Select connections by resource type:
  • Databases
  • Servers
  • Kubernetes clusters
  • Applications

​
Filter by Tags

Use tags to group related connections:
  • environment:production
  • region:us-east
  • team:platform
Type to search connections by name.

​
Select All Matching

After filtering, click Select All to add all visible connections.

​
Access Requests in Parallel Mode

If connections require access request approval:
  1. JIT Access Requests: You’ll be prompted to request access for each connection that requires it
  2. Action Access Requests: Each command submission creates approval requests
  3. Mixed: Some connections may execute immediately while others wait for approval
The Execution Summary shows which connections are waiting for approval.

​
Runbooks in Parallel Mode

You can also run Runbooks across multiple connections:
  1. Open Runbooks in the sidebar
  2. Select a runbook
  3. Enable Parallel Mode
  4. Select target connections
  5. Fill in runbook parameters
  6. Execute
All connections receive the same parameter values.

​
Best Practices

Test First

Run commands on one connection before scaling to many

Use Tags

Tag connections for easy group selection

Start Small

Begin with a few connections, then expand

Review Carefully

Check all results, not just the first one

​
Safety Tips

  1. Dry run first: If your command supports it (e.g., --dry-run), test before executing
  2. Limit scope: Don’t select all connections unless you’re sure
  3. Check the list: Review selected connections before executing
  4. Monitor results: Watch for failures as they occur

​
Troubleshooting

​
Some Connections Fail

If some connections succeed but others fail:
  1. Click the failed connection to see the error
  2. Common causes:
    • Connection is offline
    • Different OS or command not available
    • Permission issues on that specific host
  3. Fix the issue and re-run on just the failed connections

​
Command Timeout

Long-running commands may timeout:
  1. Check the command on a single connection first
  2. Consider breaking into smaller operations
  3. For very long operations, use a runbook with proper timeout settings

​
Access Request Delays

If approvals are slow:
  1. Check that approvers are online
  2. Verify Slack/Teams notifications are working
  3. Consider pre-requesting access before running parallel commands

​
Limitations

Current beta limitations:
LimitationDetails
Maximum connections50 connections per parallel execution
Command size64KB maximum command length
Result size1MB maximum output per connection
Timeout5 minutes default per connection

​
Next Steps

Runbooks

Create reusable command templates

Access Requests

Configure approval workflows

Session Recording

Audit parallel executions

Web App Overview

Explore the Web Terminal