Data Sources and Imports
Akamas Insights collects metrics from your observability tool to generate recommendations for applications running in your Kubernetes clusters. This guide covers how to configure data source integrations, start data imports, and select the active import for analysis.
Supported Data Sources
Akamas Insights integrates with the most popular observability platforms such as Dynatrace, Datadog, Prometheus, Grafana Cloud and Amazon Managed Prometheus.
For detailed configuration instructions, see:
Prometheus Configuration (includes Amazon Managed Prometheus)
If your observability platform of choice is not available, make a request at [email protected]
Configuring Data Source Integrations
The Data Sources page provides a centralized view of all configured monitoring platform integrations. Here you can review your current data source connections, test connectivity, modify configuration settings, roll API tokens, and remove integrations that are no longer needed.
Each integration displays its connection status, endpoint details, and last successful connection time. The page ensures you maintain secure, up-to-date connections to your observability platforms.
To setup a new data source head to the datasources page.
Edit: Click the edit icon to modify connection details
Delete: Remove integrations that are no longer needed
Test: Re-test connections if you update credentials
Starting Data Imports
A data import is a collection of metrics extracted from your monitoring data source for a specific time period. Each import represents a complete dataset about your cluster that Akamas Insights can analyze to generate recommendations. You can create multiple imports over time to track how your cluster evolves and measure optimization progress.
Creating a New Import
Navigate to Integrations > Data Imports
Click Start New Import
The import dialog opens with configuration options
Advanced Options
In most cases the default configuration works fine, but there are situations where you need to modify the scope of the data import. By modifying the advanced options you can focus the extraction on a specific layer, tweak how requests are made to your observability platform to handle non-standard rate limiting, or increase the granularity to capture short-lived workloads.
Click Advanced Options to customize extraction parameters:
Time Resolution: Granularity of metric data points
Options: 5 minutes, 15 minutes, 30 minutes, 1 hour, 1 day
Default: 1 hour
Lower resolution provides more detailed analysis but longer extraction time
Extraction Scope: What data to extract
Complete: Full metrics including workload and application runtime data
Infrastructure Only: Cluster and node metrics only (faster extraction)
Import Constraints
Concurrent Imports: At most 3 data imports can be run concurrently, to increase this limit contact [email protected]
Monitoring Import Progress
Data import execution time depends on many factors such as the number of monitored entities in your cluster, the time granularity, the span of the extracted time window, and the performance of your observability platform. In our experience, it might range from a few minutes to a few hours.
The data imports page provide you information on the extraction process while it is running.
Understanding Import Status
Imports can have the following statuses:
Running: Import is actively extracting data, no action needed
Paused: Import was temporarily stopped. You can resume it at will.
Completed: Import finished successfully, no action needed
Failed: Import encountered an error, please review the error or contact support
Managing Running Imports
Use the Pause, Resume, Stop, or Delete buttons to manage your imports. Pausing allows you to temporarily halt an import and resume later without losing progress. Stopping permanently cancels an import, while deleting removes completed or failed imports from the list along with the data.
Warning: Deleting a completed import permanently removes all extracted metrics from Akamas Insights. This data cannot be recovered. Before deleting, ensure you no longer need the import for analysis or comparison with future imports.
Selecting the Active Extraction
After completing one or more imports, you must select which import to use for analysis.
All analysis pages operate on the selected import. By default the latest executed import is automatically selected for analysis. To select an import for analysis you can click the select button in the data imports table. You can also review the import details by clicking on the import name before selecting it.
Note: Switching the active extraction updates all analysis pages immediately, including the dashboard, cluster details, and recommendations.
Best Practices
Data imports are quite straightforward but adhering to a few best practices can make you data import process flawless and easier to manage.
Scheduling Regular Imports
Import data regularly (e.g., weekly or monthly) to track optimization progress
Use consistent naming conventions (e.g., "Production-2025-01")
Delete old extractions to keep the list manageable and avoid losing focus on the most recent findings
Optimizing Import Performance
Start with the default time range and granularity for initial imports. If this does not work for your environment, reduce the time range (e.g., to 14 days)
Schedule imports during off-peak hours if your monitoring platform experiences heavy load
If you encounter long import times or errors, try reducing the number of clusters imported in a single data import operation
Last updated
Was this helpful?