Quick start guide for DevOps

Before you begin

Before you proceed, make sure that the following tasks are complete:

• Your IT admin installed Instana agents on your hosts and clusters. If they didn't, ask them to complete the IT Admin quick start guide first.
• You have a user account in Instana with at least User role permissions.
• You have your Instana tenant URL. It looks like https://your-company.instana.io.

Log in and explore the Instana UI

Log in to Instana and familiarize yourself with the key areas of the UI that are relevant to DevOps work.

1. Open your browser and go to your Instana tenant URL.
2. Log in with your credentials.
3. On the Home page, view a summary of your system status.
4. See the following areas that are most relevant to your work as a DevOps engineer:
   • Events > Incidents: Active problems that require attention
   • Events > Changes: Recent deployments and configuration changes
   • Applications > Services: All discovered services with their error rates and latency
   • Analytics: Query interface for unsampled, high-cardinality data
   • Automation: Automated actions and policies for incident response

Monitor your services after a deployment

Deployments are one of the most common sources of production incidents. Instana automatically correlates your deployment events with service performance data, so you can quickly spot whether a release introduced errors or regressions and act before users are affected.

1. Go to Applications > Services and select the service that you deployed.
2. On the Summary tab, review the following key metrics:
   • Calls per second: Shows the request rate and total call volume
   • Erroneous call rate: An increase indicates that the deployment introduced errors
   • Mean latency: An increase indicates a performance regression. The 90th percentile value shows typical user experience.
3. Review the detailed charts:
   • Calls: View HTTP status codes to identify error patterns (1XX, 2XX, 3XX, 4XX, 5XX). Use the Call count tab to see the volume of calls over time.
   • Erroneous call rate: Track error trends over time.
   • Latency: Compare percentiles (50th, 90th, 95th, 99th) to understand the performance distribution. Toggle between Over time and Distribution tabs to view latency trends or the distribution of response times.
4. Check Infrastructure issues & changes and Top endpoints sections for more context.
5. If you see a problem, click the Flow tab to trace requests through your service dependencies.

For more information, see Application perspectives.

Set up Smart Alerts for anomaly detection

Smart Alerts automatically detect unusual behavior in your services without requiring you to manually configure thresholds. Instana learns the normal behavior of your services and alerts you when something deviates from the baseline.

1. Go to Events and click the Smart Alert configurations tab.
2. Click Create Smart Alert.
3. Configure the Smart Alert by completing the wizard:
   • Trigger: Select what to monitor (slow calls, erroneous calls, HTTP status codes, or throughput)
   • Scope: Define where the condition applies (application perspective, service, or endpoint)
   • Threshold: Choose static (fixed value) or adaptive (automatically adjusts to metric changes)
   • Time threshold: Set when to alert (persistence over time, frequency in a time frame, or trace impact)
   • Alert channels: Configure notification channels (email, Slack, PagerDuty, or webhook)
   • Properties: Customize alert title, description, and an option to trigger an incident
   • Payloads: Add custom payloads for advanced alert routing (optional)
4. Click Create to activate the Smart Alert.

For more information, see Events and alerts.

Set up automated incident response

With Instana Automation, you can define actions that run automatically when an event or incident occurs so that you can mitigate incidents without waiting for manual intervention.

1. Go to Automation.
2. Click Create action.
3. Select an action type. Common options for DevOps include:
   • Documentation link: Access relevant documentation to diagnose or remediate a known issue directly from an event context.
   • GitHub issue: Open, close, or add comments to a GitHub issue by using the issue action sensor on the Instana agent.
   • GitLab issue: Open, close, or add comments to a GitLab issue by using the issue action sensor on the Instana agent.
   • HTTP: Add HTTP calls to call webhooks or other REST APIs by using the HTTP action sensor on the Instana agent.
   • Jira issue: Open, close, or add comments to a Jira issue by using the Instana agent's issue action sensor.
   • Manual: Provide manual instructions or steps for diagnosing or remediating an issue.
   • Script: Run automation scripts by using the script action sensor on the Instana agent.
4. Configure the action parameters for the selected action type.
5. Click Save.
6. To associate the action with an alert, go to Automation > Policies.
   1. Click Create policy.
   2. Select Event as the trigger type.
   3. Choose your Smart Alert or event definition.
   4. Select the action that you created.

Example automation scenarios:

• Restart service: Automatically restart a service when it becomes unresponsive.
• Scale resources: Trigger auto-scaling when resource usage exceeds thresholds.
• Clear cache: Clear application cache when error rates spike.
• Notify on-call: Page the on-call engineer when a critical service fails.

For more information, see Automation.

Analyze unsampled data with Unbounded Analytics

Instana collects every request, trace, and log message without sampling, so you can query the full dataset to find the exact call that caused a problem.

1. Go to Analytics.
2. From the Analytics drop-down menu, select the data type that you want to query:
   • Logs: Log messages from all services and infrastructure, with a Log console view option.
   • Applications > Calls: Individual requests across all services.
   • Applications > Traces: Full request paths across microservices.
   • Applications > Subtraces: Segments of distributed traces.
   • Websites: Website monitoring data that includes Page loads, Page transitions, Resources, HTTP requests, JS errors, and Custom events.
   • Mobile apps: Mobile application monitoring data that includes Session starts, View transitions, HTTP requests, Custom events, Crashes, Performance, and Dropped beacons.
   • Infrastructure: Infrastructure entities and custom entities.
   • Profiles: Profiling data for performance analysis.
3. Use the Add filter button to narrow down the data. The filter page provides the following options:
   • A search field to find specific filter criteria
   • Commonly Used filters relevant to the selected data type
   • Boolean operators (AND / OR) to combine multiple filters
4. For Calls and Traces, the following filters are useful for DevOps:
   • Application Name: Filter by specific application
   • Call Erroneous: Boolean flag that indicates whether the call resulted in an error
   • Call Type: Primary classification of a call based on its communication protocol or interaction pattern (for example, HTTP, DATABASE, RPC, BATCH, MESSAGING)
   • Endpoint Name: Human-readable identifier for an endpoint
   • Service Name: Identifier for the monitored service, application, or component within a distributed system
5. Use the Group option to aggregate results by specific dimensions.
6. Toggle Fast query mode for optimized performance on large datasets.
7. Click a result to see the full trace or log entry with all associated metadata.

Use case: After a deployment, query all calls to your service in the 10 minutes after deployment to identify any new errors or performance regressions.

For more information, see Unbounded Analytics.

Track deployments with pipeline feedback

Pipeline feedback automatically correlates deployment events with application and infrastructure performance data, helping you understand the impact of each release.

To create release markers from your CI/CD pipeline, use either of the following options:

• Use the Releases API in your deployment script to create release markers programmatically.
• Use one of the supported integrations, such as CircleCI, Concourse, GitHub Actions, Harness, Jenkins, or BOSH deployments.

Release markers appear on all charts, making it easy to correlate performance changes with deployments.

For more information, see Pipeline feedback.

Next steps

After you complete the initial setup, consider the following next steps:

• Review incidents weekly: Review past incidents and identify patterns that you can prevent with more automation or alerts. Go to Events > Incidents. For more details, see Analyzing incidents.
• Implement configuration management: Use Git-based configuration management to centrally manage agent configurations across your environment. For more details, see Managing agent configuration using Git.
• Automate with Terraform: Use Terraform to manage Instana configurations as code for consistent deployments. For more details, see Instana Terraform provider.
• Explore advanced integrations (if needed): For specialized use cases, explore AWS metric streams, agent-based integrations, or integration package management. For more details, see Configuring and monitoring supported technologies.

Quick reference: Key locations for DevOps tasks