Oracle Exadata

Oracle Exadata extension connects to your Exadata devices and collects Exadata characteristics published by the Oracle Prometheus exporter. Metrics collected are continuously analyzed on Dynatrace platform, leveraging topological Exadata model that the extension maintains. Unified analysis screens offer insight into health and performance of the Exadata platform, including storage server and its components, such as flash cache, cell disks, and grid disks. Out of the box alerts notify on critical issues with the Exadata resources, including memory, CPU and I/O latency.

This extension is intended for users, who:

• Look to reliably monitor Exadata components and resource utilization with comprehensive dashboards and screens
• Want to take pre-emptive measures to avoid service degradations rooting down to the Exadata resources oversubscription
• Want to leverage rich observability signals provided out-of-the-box by the Exadata Prometheus exporters
• Use Dynatrace for infrastructure monitoring and look to extend Dynatrace insights into the Exadata tier of their end-to-end infrastructure

• Integrated operational monitoring of the infrastructure, including Exadata
• Early alerting on anomalies mounting in Exadata resource usage patterns
• Root cause analysis of the Exadata bottlenecks, in the light of the entire application and infrastructure landscape observed by Dynatrace

• Follow the instructions for Using Real-Time Insight in the Oracle Exadata System Software User's Guide to configure metric stream for each type of server you want to monitor. Use the following script in DBMCLI on Database Servers (including VMs) and CELLCLI on Storage Servers to create a user with the appropriate privileges to consume the metrics.

CREATE ROLE realtime_collector
GRANT PRIVILEGE LIST ON METRICSTREAM ALL ATTRIBUTES WITH ALL OPTIONS TO ROLE realtime_collector
CREATE USER realtime_collector PASSWORD="password" ### Don't forget to change this password
GRANT ROLE realtime_collector to user realtime_collector

• Activate the extension in your Dynatrace environment web-UI using the in-product Hub and choose Monitor remotely.
• Provide the required Prometheus exporter endpoint configuration:
  • Endpoint for storage servers: https://server-name/metricstream/list?stream=true
  • Endpoint for database servers using port 7879: https://server-name:7879/metricstream/list?stream=true
  • Use Basic authentication with the user realtime_collector and password set in DBMCLI and CELLCLI

The extension package contains:

• Python scripts that connect to the Exadata endpoints and scrape the metrics
• Topology and relationship definitions for Exadata entities
• A dashboard offering high-level Exadata platform overview
• Unified Analysis pages for each Exadata entity

Summary of the entities monitored:

The Oracle Exadata extension collects metrics for:

• Exadata Server
• Exadata DB Server Filesystem
• Exadata Cell Filesystem
• Exadata Net Interface
• Exadata PMEM Cache
• Exadata Disk
• Exadata Grid Disk
• Exadata Cell Disk
• Exadata Cell
• Exadata DB Server
• Exadata IORM Database
• Exadata IORM Pluggable Database
• Exadata IORM pluggable database on cell
• Exadata IORM Consumer Group
• Exadata IORM Cluster
• Exadata Flash Cache
• Exadata Flash Cache on cell
• Exadata Flash Cache on DB
• Exadata Flash Cache on PDB

• Network access from a Dynatrace ActiveGate to the Exadata endpoints specified in the Get started section
• Oracle Exadata System Software 22.1.0+
  • This Extension uses API introduced in Exadata System Software 22.1.0
  • Your Exadata Database Machine and Exadata Storage Server would carry the same version number as the Exadata System Software indicates
• Note that this extension does not support ExaCC

Q: Can I monitor ExaCC with this extension?

A: This extension does not support monitoring ExaCC (Exadata Cloud at Customer). We are working on adding ExaCC monitoring support in the future, but not in this extension due to API differences between regular Exadata and ExaCC.

Q: How to comprehend Exadata metric names, when theres so many of them?

A: Exadata metric keys have a self-explanatory structure. Knowing it, you will always be able to understand what specific metric represents, exactly, even if its name or description looks similar to other available metrics.

Key to deciphering Exadata metric names:

The value of the name attribute is a composite of abbreviations. The attribute value starts with an abbreviation of the object type on which the metric is defined:

CD_ (cell disk)
CG_ (IORM consumer group, database-qualified)
CL_ (cell)
CT_ (IORM category)
DB_ (IORM database-level)
FC_ (flash cache)
FL_ (flash log)
GD_ (grid disk)
IORM (Input Output Resource Management)
N_ (network, IBPORT, HOST_INTERCONNECT)
PDB_ (IORM pluggable database)
SIO_ (Smart IO)

After the abbreviation of the object type, most of the name attributes contain one of the following combinations to identify the operation:

IO_BY (I/O amount)
IO_RQ (number of I/O requests)
IO_TM (I/O latency)
IO_WT (I/O wait time)
FC_IO_BY (Flash cache I/O amount)
FC_IO_RQ (Flash cache I/O requests)
FD_IO_BY (Flash disk I/O amount)
FD_IO_RQ (Flash disk I/O requests)
FD_IO_TM (Flash disk latency)
FD_IO_UTIL (Flash disk utilization percentage)

Next, in the name could be

_R (for read) or 
_W (for write). 

Following that in the name attribute value there might be

_SM or _LG to identify small or
_LG to identify large blocks, respectively. 

At the end of the name, there could be

_SEC to signify per seconds or 
_RQ to signify per request.

For consumer group and category metrics, read or write details are omitted.

For example: CD_IO_RQ_R_SM is the number of requests to read small blocks on a cell disk. GD_IO_TM_W_LG is the microseconds of I/O latency writing large blocks on a grid disk.

Source: Exadata Database Machine User's Guide section DESCRIBE METRICDEFINITION.

Q: Do I have to configure metric stream on all DB servers and Cell servers, or one of them is enough?

A: You need to configure metric streams on each Exadata server you intend to monitor. Use DBMCLI utility on DB servers. Use CELLCLI utility on cell servers. Or use utilities designed to run administration commands across a number of servers (dcli, exadcli) - see Exadata Software Overview.

Q: How to set up and verify on the Exadata side that metrics are available?

A: Metrics must be enabled for Real-Time Insight on the Exadata side. Most likely, metrics are enabled by default on your Exadata system. If they are not, enable metrics for Real-Time Insight with the command: alter metricdefinition finegrained=enabled where finegrained=disable You can check what metrics are enabled using command: list metricdefinition where finegrained=enabled detail The command returns a detailed list of metrics that are enabled for Real-Time Insight.

Q: What is the DDU Consumption of this extension?

A: The formula for DDU consumption of the extension is:

( ( 19 * the number of servers )
+ ( 10 * the number of servers * the number of disks on the server )
+ ( 6  * the number of servers * the number of network interfaces on the server )
+ ( 12 * the number of database servers )
+ ( 1  * the number of database servers * the number of file systems on the DB server )
+ ( 1  * the number of cell servers * the number of file systems on the cell server )
+ ( 25 * the number of cell servers * the number of disks on the cell server )
+ ( 99 * the number of cell servers )
+ ( 16 * the number of cell servers * the number of databases on the cell server )
+ ( 16 * the number of cell servers * the number of pluggable databases on the cell server )
+ ( 5 * the number of cell servers * the number of DB/PDB * the number of consumer groups on the DB/PDB )
+ ( 2 * the number of category plans * the number of cell servers )
+ ( 2 * the number of IP addresses in queue )
+ ( 7 * the number of servers * (4 * the number of ROCE interfaces on the server) )
+ ( 6 * the number of cell servers * (2 * the number of cell disks on the cell server) )
)* 525.6 DDUs/year

Note that this accounts for all extension feature sets enabled, i.e. all possible metrics collected. In case of license consumption concerns, you may disable optional feature sets.

DDU cost above does not include any possible Log events or Custom events triggered by the extension. For more information on this, please visit the DDU log event cost and DDU custom event cost pages.