Observability & Management services to manage a large Oracle Database fleet

Implementing an observability strategy requires collecting IT environment data such as metrics, logs, events, or traces.  Good, solid Observability means using your data and assets wisely so that you can easily predict outages, excessive use of resources, or applications’ poor performance before they happen.

Being proactive is a must in a distributed environment and saving resources is directly correlated with your budget.  OCI Observability & Management (O&M) provides a strategy to make day-to-day activity easier for both DBAs and IT managers. DBAs should be able to provision PDBs, plug/unplug, and create test and dev environments with a click of a button. IT Managers should be able to provide a forecast of the budget based on current resource utilization and trend analysis.

O&M services improve the communication between teams and departments by offering a single pane of glass options to optimize the interaction. It also makes the resources needed evident by correlating costs for each.  Unlike basic monitoring tools, O&M solutions provide the ability to forecast resource utilization and notify customers days before a resource outage. Here is an example of forecasting of CPU usage, across all databases.

For Engineered Systems environments, you can reduce outages by clustering events and data by similarity, identifying critical patterns, and comparing the trend to prevent error propagations.

In this blog, we explain how to leverage the different Oracle tools and services for your O&M needs, specifically for database environments. 

Typical database environments to leverage O&M tools

Observability for on-premises/hybrid

In a hybrid scenario, some databases are on-premises, and some in OCI.

In a hybrid environment, with on-premises and Cloud databases, Enterprise Manager (EM) is the optimal solution for these reasons:

• EM is on-premises so it makes the discovery of targets easy
• EM has a rich interface to monitor and administer database targets
• Benefit from the EM topology without discovering it from scratch. They have been already discovered in the Enterprise manager logs

Consider enriching your Observability platform by extending EM with OCI Ops Insights and Logging Analytics services capabilities, such as:

• Resource Usage Forecast for budget allocation
• Utilization trend by department/team for chargeback
• Resource utilization limits early warning
• SQL degradation analysis
• Database Alerts and trace log management for root cause analysis

Choose to deploy EM on an OCI Compute VM and use the full Stack provided in OCI Marketplace.  Similar to an on-premise EM, the OMS and Repository maintenance is under your control.

Observability for ExaCC

Everything is in your data center, similar to the example shown below:

In this scenario, database management is limited to the database layer.  However, EM capabilities can be enriched  with OCI Op Insights, Exadata Insights, and Logging Analytics services for the following benefits:

• Resource Usage Forecast for budget allocation
• Utilization trend by VM cluster/department/team for chargeback
• Resource utilization limits early warning
• SQL degradation analysis
• Database Alert and trace log management for root cause analysis (suggest importing application logs)

Observability for OCI ExaCS/DBCS

In this case, everything runs in OCI.

Use the native OCI services to save on maintenance for EM components.

Native OCI services like Database Management and Ops Insights are agentless (creation without an agent) for OCI-based Oracle databases. The Logging Analytics service can be used with Management Agent for security reasons and will allow you to ingest any file or data from database tables in addition to pre-defined logs sources.

Provisioning and database resource management 

Provisioning involves resource creation, termination, definition, assignment, and all other activities that allow the creation and use of a database. This also includes setting resource limits in case various teams share resources.  Here are various solutions used for provisioning needs:

Real Time Monitoring

Real-time monitoring means continuously delivering updated data about systems, processes, or events. Such monitoring provides information streaming at zero or low latency, there is minimal delay between data collection and analysis. It enables quick detection of anomalies, performance issues, and critical events.

Oracle provides the following products/services for real-time monitoring:

Performance and Tuning solutions to evaluate database systems

Performance diagnostics and tuning are critical for any IT environment. For database systems, this activity can be proactive or reactive.  For example, reacting to an increased workload is critical to tune the subsystem as it can end up in an outage. Performance tuning can be at the system resource level (CPU/memory/storage utilization) or a higher level like SQL statements response time. It is important to analyze and correlate the information to correctly identify the bottleneck.  Here are Oracle solutions to help with performance evaluation and tuning:

Database fleet administration solutions

Database Administrators are responsible for the management and maintenance of company databases. Their duties include maintaining adherence to a data management policy and ensuring that these essential pieces of equipment are functional. Activities include instance start and stop, database backup and restore, key management, and resource assignment up to a fixed storage level.

The following Oracle tools and solutions are key for a database fleet admin:

Patching a database fleet safely

Patching is one of the important stages of the product lifecycle. It enables you to keep the software product updated with bug fixes. Oracle releases several types of patches periodically to maintain a secure and bug-free stack. Although OCI simplifies it, patching has always been a challenging phase of the lifecycle because it is complex, time-consuming, and can involve downtime. These Oracle tools/services can reduce the risk and allow database admins to safely patch their database fleet.

Cost control and chargeback provide visibility into IT services

Cost control is the practice of identifying and reducing business expenses to increase profits, and it starts with the budgeting process. Cost control is an important factor in maintaining and growing profitability.  IT chargeback can provide greater visibility into the costs of IT services and infrastructure usage, enabling organizations to identify opportunities for cost optimization and reduce wasteful spending. Cost Control and chargeback are important topics for the company that adopts the Cloud or a new FinOps challenge. In this scenario, savings in consumption is directly connected to the business.

Troubleshooting to identify root cause and prevent outages

Database issues can happen on several levels.  To identify problem root causes, it is important to be able to correlate resources, drill down into the issues, and analyze trends in the systems. The root cause can be an application so it is important to also have visibility into this information. Troubleshooting also helps avoid error propagation which is why it is important to notice the issues as early as possible.

All these Oracle tools are important for troubleshooting and preventing outages:

Oracle offers several tools to manage and monitor large database fleets. It is important to extend EM with O&M capabilities with diagnostic tools like forecast or pattern identification offered in OCI. Explore more and perform Hands-on Lab activities on Oracle Cloud-for free, Use the Oracle Cloud Free Tier Account. 

Resources:

• Integrating Enterprise Manager with OCI Services
• Oracle Enterprise Manager Cloud Control for Oracle Exadata Cloud
• Enterprise Manager
• Observability & Management 
• Oracle Observability and Management blogs 
• Oracle named a leader in hybrid cloud management 
• Day One and Beyond:  Introduction to Observability & Management video