Workflow for Monitoring Integrations

Initial Monitoring Tasks

Before an integration goes live in production, optimize your monitoring workflow by completing a few tasks.

You might need to complete some tasks again when your organization prepares to release a major update to an integration.

Task	More information
Complete performance testing	See Measure an Integration's Performance.
Complete load testing	Oracle recommends checking the scalability of a new integration by performing load testing. Checking the performance of an integration under different loads helps you understand how the integration will perform in a real-world scenario. For instance, how will your order creation integration handle Black Friday loads? After testing the integration under load, measure its performance. See Measure an Integration's Performance.
Follow your organization's standard operating procedures for preparing for an integration to go live	For example, document the following information in a location that stakeholders can access: Name of the integration Overview of the business process that the integration automates Applications that are involved Contact names, including backups, for who can help with failures and who should be informed in the event of a failure Name of the connectivity agent that the integration uses, if any See View the Dashboard.
If your organization maintains custom monitoring tools, determine whether the new integration requires updates to the tools	See About Building Your Own Observability Tools.
For schedule integrations, make sure the start time doesn't conflict with other integrations	For example, don't run twenty schedule integrations at the close of business. Instead, stagger the integrations by at least a few minutes. See View the Calendar of Schedule Integration Runs.
Set up notifications	Notifications provide daily or hourly emails about failures or the overall health of all your integrations. See Send System Status Reports with Notification Emails.
Determine when responsibility for the integration changes hands	If you're responsible for monitoring an integration that you built, this task doesn't apply to you. However, if you're responsible for monitoring someone else's integration, work with the integration developer to determine a hand-off schedule. For instance, do you take ownership when the integration goes live, or only after the integration has run successfully in production for a period of time? Setting expectations from the beginning helps ensure the successful launch of a new business process automation.
(When your organization creates a new version of a live integration) Verify that the new version doesn't contain any performance regressions	After you version and update an integration, verify that the new version of the integration performs as expected. For example, you can compare the average running times of the old and new integrations. See Determine an Integration's Typical Processing Times.

Daily Monitoring Tasks

Your organization's requirements how often you perform these monitoring tasks. Typically, organizations perform these tasks at least daily, often at predetermined times each day.

Task	More information
Determine whether any integrations experienced errors	See Find Integrations That Experienced Errors.
Determine whether any integrations timed out or were aborted (canceled)	See Find Integrations That Timed Out or Were Aborted.
Determine whether an integration has experienced performance issues	Troubleshoot slow-running integrations and verify that integrations meet your organization's performance requirements. See Measure an Integration's Performance.
Identify the most recent updates to an integration	Consider a scenario in which an integration that has previously run as expected is suddenly not working as expected. An inadvertent update could be to blame. Check the integration's audit history to understand who updated it, what changes they made, and when they updated it. For instance, if unscheduled changes occurred, discuss the update with the developer and determine whether the changes need to be rolled back. See Check the Audit History for an Integration or Other Component.
Determine why fewer integration instances are running than expected	See Fewer Integration Instances Ran Than Usual.
Determine whether all connectivity agents are available	If a connectivity agent isn't available, any integrations that use the agent fail. Inform the administrator who is responsible for the connectivity agent if any agents go down. See Monitor Agents.
Identify the connections that cause the most errors	See Find the Connections with the Most Errors.

Occasional Monitoring Tasks

Routine analysis work supports operational excellence. Use your organization's operating procedures to determine how often you complete these tasks. For example, you might complete some tasks weekly, monthly, or quarterly.

Task	More information
Verify that integrations continue meeting your organization's performance requirements	See Measure an Integration's Performance.
Ensure that the tracing level is set correctly for integrations	See Check an Integration's Tracing Level.
Make sure that your organization's active integrations stay below the limit	On the Dashboards page, see the numbers in the Active integrations box. To learn more about the service limits, Service Limits in Provisioning and Administering Oracle Integration 3
Confirm that your schedule integrations start at staggered times	If too many schedule integrations start at the same time, they might experience contention. Ease the contention by staggering the start times. See View the Calendar of Schedule Integration Runs.
See a snapshot of the current status of the components your organization has developed	See See the Current Status of All Integration Components.
Update the information you've documented for integrations	For example, if you document the contacts and stakeholders for an integration, periodically review the information and make updates for organizational changes and employee turnover.