[role="xpack"]
[[service-overview]]
=== Service overview

Selecting a <<services,*service*>> brings you to the *Service overview*.
The *Service overview* contains a wide variety of charts and tables that provide
visibility into how a service performs across your infrastructure.

[discrete]
[[service-latency]]
=== Latency

Response times for the service. You can filter the *Latency* chart to display the average,
95th, or 99th percentile latency times for the service.

[role="screenshot"]
image::apm/images/latency.png[Service latency]

[discrete]
[[service-throughput-transactions]]
=== Throughput and transactions

The *Throughput* chart visualizes the average number of transactions per minute for the selected service.

The *Transactions* table displays a list of _transaction groups_ for the
selected service and includes the latency, traffic, error rate, and the impact for each transaction.
Transactions that share the same name are grouped, and only one entry is displayed for each group.

By default, transaction groups are sorted by _Impact_ to show the most used and slowest endpoints in your
service. If there is a particular endpoint you are interested in, click *View transactions* to view a
list of similar transactions on the <<transactions, transactions overview>> page.

[role="screenshot"]
image::apm/images/traffic-transactions.png[Traffic and transactions]

[discrete]
[[service-error-rates]]
=== Error rate and errors

The *Error rate* chart displays the average error rates relating to the service, within a specific time range.

The *Errors* table provides a high-level view of each error message when it first and last occurred,
along with the total number of occurrences. This makes it very easy to quickly see which errors affect
your services and take actions to rectify them. To do so, click *View errors*.

[role="screenshot"]
image::apm/images/error-rate.png[Error rate and errors]

[discrete]
[[service-span-duration]]
=== Span types average duration and dependencies

The *Average duration by span type* chart visualizes each span type's average duration and helps you determine
which spans could be slowing down transactions. The "app" label displayed under the
chart indicates that something was happening within the application. This could signal that the
agent does not have auto-instrumentation for whatever was happening during that time or that the time was spent in the
application code and not in database or external requests.

The *Dependencies* table displays a list of downstream services or external connections relevant
to the service at the selected time range. The table displays latency, traffic, error rate, and the impact of
each dependency. By default, dependencies are sorted by _Impact_ to show the most used and the slowest dependency.
If there is a particular dependency you are interested in, click *View service map* to view the related
<<service-maps, service map>>.

IMPORTANT: A known issue prevents Real User Monitoring (RUM) dependencies from being shown in the
*Dependencies* table. We are working on a fix for this issue.

[role="screenshot"]
image::apm/images/spans-dependencies.png[Span type duration and dependencies]

[discrete]
[[service-instances]]
=== All instances

The *All instances* table displays a list of all the available service instances within the selected time range.
Depending on how the service runs, the instance could be a host or a container. The table displays latency, traffic,
errors, CPU usage, and memory usage for each instance. By default, instances are sorted by _Throughput_.

[role="screenshot"]
image::apm/images/all-instances.png[All instances]

[discrete]
[[service-metadata]]
=== Service metadata

To view metadata relating to the service agent, and if relevant, the container and cloud provider,
click on each icon located at the top of the page beside the service name.

[role="screenshot"]
image::apm/images/metadata-icons.png[Service metadata]

*Service information*

* Service version
* Runtime name and version
* Framework name
* Agent name and version

*Container information*

* Operating system
* Containerized - Yes or no.
* Total number of instances
* Orchestration

*Cloud provider information*

* Cloud provider
* Availability zones
* Machine types
* Project ID