Glossary for Splunk Observability Cloud 🔗
- agent (deployment method) 🔗
An agent is a deployment method where an instance of the Splunk Distribution of OpenTelemetry Collector runs with the application or on the same host as the application. For example, when you configure the Splunk Distribution of OpenTelemetry Collector for Linux, Kubernetes, or Windows, you are using the agent deployment method.
- alert 🔗
An alert is triggered when the conditions for a detector rule are met. For example, a detector monitoring the number of requests served by an application may have a rule that produces an alert if the number is below a static threshold, for example, 20 requests per minute, and/or above a calculated one, for example, the mean + 3 standard deviations of the number of requests per minute over the past hour.
- analytics 🔗
Analytics are the mathematical functions that can be applied to a collection of data points. For a full list of analytics that can be applied in Splunk Infrastructure Monitoring, see the Functions reference for Splunk Observability Cloud.
- call stack 🔗
A call stack is the data structure used by a machine to keep track of which methods are currently being called. When the active call stack is sampled, the result is a stack trace.
- counter metric 🔗
The counter metric type represents data that is a count of occurrences in a time interval. It measures occurrences of an activity or event, for example, the number of web pages served by a website or the number of exceptions in a process. Summing counters over a period of time produces the net activity in that interval. Counters can only take integer values of zero or greater and are reset to zero at the conclusion of each reporting interval.
- cumulative counter metric 🔗
The cumulative counter metric type represents a running count of occurrences. It typically represents the total activity in the lifetime of an application or process. Cumulative counters are NOT reset with each reporting interval. Examples of cumulative counters include the total number of Splunk Infrastructure Monitoring API calls served since a web server started up, or the total number of bytes transmitted by an interface since it started up. Cumulative counters can also be used to derive incremental values, just as counters can.
- detector 🔗
A detector monitors a signal for conditions or issues that you care about.
Those conditions or issues are expressed as one or more rules that trigger an alert when the conditions in the rules are met. Individual rules in a detector are labeled according to criticality: Info, Warning, Minor, Major, and Critical.
For example, a detector that monitors the latency of an API call may go into a critical state when the latency is significantly higher than normal, as defined in the detector rules.
For more information, see Introduction to alerts and detectors in Splunk Observability Cloud.
- dimension 🔗
A dimension is a key/value pair that, along with the metric name, is part of the identity of a time series. You can filter and aggregate time series by those dimensions across Infrastructure Monitoring.
- event 🔗
An event is a periodic occurrence that can be represented as a structured log line to Splunk Infrastructure Monitoring. For example, the values could be represented as any combination of key/value pairs. Events are secondary to metrics in Infrastructure Monitoring and are meant to provide context for the metric data. Events can be displayed on charts and viewed in the Events sidebar. See Add context to metrics using events.
- event time series 🔗
An event time series (ETS) is a sequence of events uniquely identified by the event name and optional, additional dimensions. For example, an event time series with the name
code pushand the dimension
repositorycan be created to record code push events for a given repository. An example of such an ETS could be
- flame graph 🔗
The flame graph is a visual representation of a collection of stack traces. For every line of code in the stack trace, there is a corresponding line in the flame graph. The width of each bar in the flame graph represents the number of times the respective line of code appears in the stack traces that have been collected in the time range of the flame graph. For example, if a line of code occupies 100% of the width of the flame graph, then that line of code appears in all of the stack traces in the collection. The y-axis of the flame graph shows the depth of the stack trace. The colors of the flame graph are random. The x-axis is not ordered by time. The ordering of stack traces from left to right is random and it does not correlate to a time-based sequence.
- flappy 🔗
A detector is said to be “flappy” when it triggers and clears alerts too frequently. For example, if you have a detector set to trigger an alert when a value reaches 90%, and the signal you are monitoring regularly spikes and dips around this value, alerts will be triggered and cleared too often to be of value. To reduce this flappiness, you might want to specify that the value must remain at 90% for a specified amount of time before triggering an alert.
- gateway (deployment method) 🔗
Gateway is a deployment method where the Splunk Distribution of OpenTelemetry Collector is running by itself. When you configure the Splunk Distribution of OpenTelemetry Collector as a standalone package, you are using the gateway deployment method.
- gauge metric 🔗
The gauge metric type represents data that has a specific value at each point in time. It measures the value of something over time. Examples of gauges used in monitoring include CPU utilization percentage, % free JVM heap, or the size of an internal queue in an application. Reporting frequency (in other words, how often you take a reading) is most important for gauges, as higher frequency is typically associated with higher accuracy.
For example, measuring CPU utilization every 5 minutes means that any peaks and valleys that may have occurred between readings are missed, and it’s entirely possible that those peaks or valleys may be significant.
- integration 🔗
An integration is a configurable component of Observability Cloud that connects Observability Cloud to a third-party service. Most integrations connect third-party data services, but Observability Cloud also offers SSO and notification integrations.
- metric 🔗
Metrics are the primary form of data you send into Splunk Infrastructure Monitoring. A metric is a periodic measurement that is represented as a numerical value. The same metric can be reported from multiple sources or emitters. Typically, each unique combination of a source and a metric results in a metric time series.
- metric time series 🔗
A metric time series (MTS) is defined by the unique combination of a metric and a set of dimensions (which may be empty). The most common dimension is a source, like a host or instance for infrastructure metrics, or an application component or service tier for application metrics. The output of analytics pipelines are also metric time series.
- MTS 🔗
See metric time series.
- muting rule 🔗
- property 🔗
Properties are key-value pairs that can be bound to metrics, dimensions, or time series. They define arbitrary text data that can be used to provide additional operational information on the objects they are associated with. Properties are different from dimensions in the sense that they do not take part in the identity of a time series; changing the value of a property does not affect the identity of that time series.
Property values are most frequently used as dynamic filters for charts (for example, show 90th percentile of CPU utilization for servers with a location property value of “Seattle”), or for groupings (for example, show 90th percentile of CPU utilization for servers, grouped by location value).
- rollup 🔗
An accumulation of data points, with some mathematical or statistical expression applied to it. For example, a 95th percentile calculation over a 1-week window. In an Infrastructure Monitoring plot, rollups determine how Infrastructure Monitoring prepares data points for use in charts or analytic computations.
For example, if you change the time range from -1m (past minute) to -1w (past week), multiple data points may be rolled up into one using a rollup function, such as Average, so the data points for the wider timeframe can be effectively displayed.
For more information, see Rollups.
- rule 🔗
A detector contains one or more rules that specify conditions under which the detector triggers an alert, the severity of the alert, and the recipients of notifications that are sent when the condition occurs and when it clears.
For more information, see Build detector rules.
- signal 🔗
In the context of a chart in Infrastructure Monitoring, a signal is the metric time series that you want to plot on a chart or use as an input to a detector or to additional analytics.
- stack trace 🔗
A stack trace is a sampled snapshot of the call stack. The stack trace contains the class name, method name, and line number in the call stack for a given thread. For example, AlwaysOn Profiling captures a stack trace for every running thread in the Java Virtual Machine. When stack traces are sampled across all VM threads, the result is a thread dump.
- span 🔗
A span is a single operation within a trace. A session is made up of a collection of spans and traces.
- Splunk Distribution of OpenTelemetry Collector 🔗
The Splunk Distribution of OpenTelemetry Collector is a package that bundles the Splunk Distribution of OpenTelemetry Collector with additional components to provide integrated collection and forwarding of traces, metrics, and logs for a specific platform. Configuring the Splunk Distribution of OpenTelemetry Collector uses the agent deployment method.
- tag 🔗
Tags can be thought of as labels or keywords assigned to dimensions, metrics, and other objects. They are not key/value pairs.
The primary use case for tags is when there is a one-to-many relationship between the tag and the object you are assigning it to. For example, suppose you have hosts that are running multiple apps. You can create a tag for each app, then apply multiple tags to each host to specify the apps that are running on that host.
- trace 🔗
A trace is a collection of operations that represents a unique transaction handled by an application and its constituent services. Traces are made of spans, which are calls that microservices make to each other.