How to Create an IT Dashboard in Looker

Building an IT dashboard in Looker gives you a real-time command center to monitor the health and performance of your entire technology stack. It transforms scattered logs and metrics into a clear, actionable view of your systems. This guide will walk you through the entire process, from planning your metrics to building an interactive dashboard step-by-step.

Why Build an IT Dashboard?

An IT dashboard centralizes critical system information, moving you from a reactive "fire-fighting" mode to proactive performance management. Instead of digging through server logs or system alerts after a problem occurs, a dashboard presents a live overview of your infrastructure. This helps you spot trends, identify potential issues early, and understand how different parts of your system interact.

Common use cases for an IT dashboard include:

• Infrastructure Monitoring: Tracking server CPU, memory, and disk utilization to prevent outages from resource exhaustion.

• Application Performance Monitoring (APM): Keeping an eye on application response times, error rates, and user transaction performance to ensure a smooth user experience.

• Network Health: Visualizing network latency and bandwidth usage to identify bottlenecks.

• Security Auditing: Monitoring for security events like failed login attempts or unusual access patterns.

• Helpdesk and Support Operations: Tracking ticket volume, resolution times, and team performance to optimize support workflows.

Before You Build: Planning Your Dashboard

The success of your dashboard depends on the groundwork you do before creating a single chart. Rushing this stage often leads to dashboards that are cluttered, confusing, or simply don't answer the right questions.

1. Define Your Audience and Key Goals

Start by asking: Who is this dashboard for? The information needed by a Chief Technology Officer (CTO) is very different from what a DevOps engineer needs. Identify your primary audience and their key objectives.

• For Executives (CTO/VP of IT): They need a high-level overview. Focus on business-centric metrics like system uptime percentage, average application response time across the board, and overall security incidents.

• For DevOps/SRE Teams: They need granular, real-time data to troubleshoot. Their dashboard should feature detailed metrics like CPU usage per server, specific application error rates, and memory leak graphs.

• For Helpdesk Managers: Their focus is on operational efficiency. Key metrics include ticket inflow vs. outflow, average time to resolution, and agent-specific performance numbers.

2. Identify Your Key Performance Indicators (KPIs)

Once you know your audience, list the specific KPIs that will help them achieve their goals. A good dashboard tells a story by focusing on a few vital metrics rather than trying to show everything. Here are some common IT KPIs to consider:

Infrastructure Metrics

• CPU Utilization (%)

• Memory Usage (%)

• Available Disk Space (GB or %)

• Server Read/Write Speed

• Uptime / Downtime (%)

Application Performance Metrics

• Average Response Time (ms)

• Apdex Score (Application Performance Index)

• Error Rate (e.g., number of 5xx errors)

• Request Rate (requests per minute)

Service Desk Metrics

• Number of Open Tickets

• First Response Time

• Average Time to Resolution

• Ticket Backlog

• Customer Satisfaction (CSAT) Score

3. Gather Your Data Sources

Where does the data for these KPIs live? An IT environment is powered by dozens of different tools, and your dashboard will need to connect to them. Common data sources include:

• Server/VM Logs: Amazon CloudWatch, Google Cloud Monitoring, Azure Monitor, or raw server log files.

• APM Tools: New Relic, Datadog, Dynatrace.

• Ticketing Systems: Jira Service Management, ServiceNow, Zendesk.

• Network Monitoring Tools: Nagios, Zabbix.

• Databases: PostgreSQL, MySQL, SQL Server where performance data might be stored.

Make a list of every source you'll need to pull data from. Gaining access and understanding the structure of each source is a crucial upfront task.

Step-by-Step: Creating Your IT Dashboard in Looker

With a solid plan in place, you’re ready to start building in Looker. The process follows a logical path from connecting data to arranging your final visualizations.

Step 1: Connect Your Data Sources

Looker works by connecting directly to your existing databases and data sources. It doesn't store your data, it queries it live.

1. Navigate to the Admin panel in Looker.

2. Under the Database section, click on Connections.

3. Click Add Connection and select the appropriate database dialect (e.g., PostgreSQL, Google BigQuery, Amazon Redshift).

4. Enter your database credentials and connection details. Ensure the database user Looker will use has the necessary read-only permissions for the data you need.

Repeat this process for each of the data sources you identified during your planning phase.

Step 2: Model Your Data with LookML

This is where Looker’s real power lies. LookML is Looker's modeling language, and it allows you to define the dimensions, measures, calculations, and relationships in your data. It acts as a single source of truth for your business logic, ensuring everyone analyzes the same metrics calculated in the same way.

1. Go to the Develop section and create a new LookML Project.

2. Within your project, you'll create Views. A view typically corresponds to a single table in your database (e.g., a server_logs view for your server performance table).

3. Inside each view file, define your dimensions and measures:

   • Dimensions are the "group by" fields - the attributes of your data. Examples include server_name, timestamp, error_type, or ticket_status.

   • Measures are the aggregate calculations you perform on your dimensions. Examples include count, sum, average, max, or min. You could create an average_cpu_usage measure or total_open_tickets count.

For example, a simple view for server CPU data might look like this:

Step 3: Create Explores to Make Data Queriable

A LookML Explore is the starting point for business users to build a report. It defines join relationships between different views, allowing you to analyze data from multiple sources in a single query.

In your LookML model file, you define your Explores. For an IT dashboard, you might create an it_operations explore that joins your server_metrics view with an application_errors view using a common key like a timestamp or server ID.

Step 4: Build Your Visualizations (Looks)

Once your model is set up, you can start exploring your data and building visualizations, which Looker calls "Looks."

Let’s build a tile for CPU Utilization Over Time:

1. Click on Explore and select your newly created server_metrics Explore.

2. In the left-hand pane, you'll see the dimensions and measures you defined in LookML.

3. Select a time dimension (e.g., Timestamp Date) and a measure (e.g., Average CPU Utilization).

4. You can add a filter if needed, like filtering for a specific Server ID.

5. Click Run. Looker generates a data table with your results.

6. Under the Visualization tab, choose a line chart. You can customize colors, axis labels, and title.

7. Once you're happy with it, click the gear icon in the top right and select Save > As a new Look. Give it a descriptive name like "CPU Utilization - Last 30 Days."

Repeat this process for every KPI you planned. Create a "Single Value" tile for current uptime, a bar chart for ticket status, and a table for the top 5 most frequent application errors.

Step 5: Assemble Your Dashboard

Now it's time to bring all your individual Looks together into a cohesive dashboard.

1. Navigate to your desired folder, click New, and select Dashboard.

2. Give your dashboard a name, like "IT Operations Command Center."

3. Click Add Tile. You can now add the Looks you saved, create new visualizations directly, or add text tiles for headings and descriptions.

4. Drag and drop the tiles to arrange them logically. Place your most critical, high-level KPIs at the top left, as this is where people naturally look first. Group related metrics together in sections like "Infrastructure Health" and "Helpdesk Performance."

Step 6: Add Interactive Filters

A static dashboard is useful, but an interactive one is empowering. Filters allow users to slice and dice the data themselves.

1. In dashboard edit mode, click Filters in the top toolbar and select Add Filter.

2. A common filter is a Date filter. Create one and label it "Date Range."

3. For each relevant tile on your dashboard, link it to the filter. Go to the tile's menu (three dots), select Edit Tile, and go to the Filters tab. Select which field on that tile (e.g., Timestamp Date) should be controlled by your new "Date Range" dashboard filter.

4. You can add other filters for server_name or app_name to allow for deep-dive analysis.

Once saved, users can now change the date range or select a specific server, and all connected tiles on the dashboard will update automatically.

Dashboard Best Practices

• Tell a story: Arrange your visuals to guide the user from a high-level overview down to the details. Start with summary metrics and move to more granular charts.

• Use color with purpose: Don't just make it colorful. Use color to signal status. For example, use conditional formatting to make single-value tiles turn red when a KPI crosses a dangerous threshold.

• Keep it clean and simple: Resist the urge to pack too much onto one screen. If a dashboard becomes too cluttered, it loses its at-a-glance value. It's better to have two or three focused dashboards than one that tries to do everything.

• Set up alerts: Use Looker's alerting feature to get notified via email or Slack when a metric hits a certain threshold (e.g., when disk space drops below 10%). This brings proactive monitoring to the next level.

Final Thoughts

Building a robust IT dashboard in Looker organizes your operational data into a powerful tool for monitoring, troubleshooting, and strategic planning. By carefully planning your metrics, modeling your data properly with LookML, and designing with your audience in mind, you create a central command center that provides invaluable insights into the health of your systems.

For many teams, the setup-intensive process and LookML learning curve associated with tools like Looker can be a significant hurdle. At Graphed , we’ve focused on simplifying this entire workflow. By connecting your data sources and using natural language to ask questions like, "Show me server uptime by device for last week," you can generate the same live visualizations in seconds. Our goal is to give you back the time you’d spend on a complex setup so you can focus on acting on the insights, not just searching for them.