Create a Container Job
Container Jobs allow you to execute scheduled workloads using container images — perfect for batch tasks, periodic data syncs, or report generation.
This guide walks you through creating your first container job using the Nexaa portal.
Prerequisites
- A Nexaa account with billing completed.
- A container image that performs the task you want to schedule (must be hosted in a public or authorized registry).
- Optional: A private registry configured in your Namespace if your image is private.
Step 1: Log In
Go to the nexaa portal and sign in using your Nexaa credentials.
Step 2: Navigate to Your Namespace
- From the dashboard, select your Namespace or create a new one.
- Go to the tab "Container Jobs".
Step 3: Create a New Job
- Click "Add Container Job".
- Fill in the following fields:
| Field | Description |
|---|---|
| Container Name | A unique identifier for your job (e.g. daily-report) |
| Container Image | The image you want to run (e.g. my-org/reporter:latest) |
| Registry Credentials | The registry source of where your image is stored |
| Resource | Select the required resources for your job |
| Environment Variables | Key-value pairs for environment variables |
| Command (optional) | Override the image's default ENTRYPOINT and Command |
| Persistent Volume | Mount an existing or new persistent volume |
| Schedule (Cron) | When the job should run (e.g. 0 12 * * * * for every day at 12:00) |
| Enable Job | Enable the job so it will be excuted at the scheduled time |
Optionally, attach secrets or volumes if your job depends on them.
Step 4: Save and Deploy
- Click "Create Job".
- Your job will now appear in the list with its next execution time.
Step 5: Monitor Job Executions
- Click on the job to view its execution history.
- For each run, you can inspect:
- Status (success/failure)
- Start/End time
- Logs output from STDOUT/STDERR
Tips & Best Practices
- Test your image manually before scheduling to avoid runtime surprises.
- Avoid writing to the container filesystem — use external storage or databases.
- Keep the job duration short to avoid overlaps (concurrency is not supported).
- Use secrets for any credentials instead of hardcoding environment variables.
- Pause Job Disable the job to keep the configuration but pause execution until enabled again.