Telemetry As A Service (TMS)
Telemetry as a Service allows collecting and storing your infrastructure metrics - CPU utilization, disk Read/Write operations and
network traffic. The service is based on Gnocchi, a metrics database platform, and collectd, a service collecting the instance metrics and
sending them to Gnocchi.
Currently, Telemetry as a Service allows gathering and storing metrics from Ubuntu16, Ubuntu14, CentOS7 based instances. Support of other images will be implemented later.
At the moment, Telemetry as a Service is supported only in EPAM regions.
For more details on the components used in the service deployment, visit the official websites of Gnocchi and
Gnocci version 4.0 and collectd version 5.7 are supported.
Related CLI Commands
The table below provides the list of service-related commands and their descriptions:
|or2-manage-service ... -s telemetry -a
||or2ms ... -s telemetry -a
||Starts the service in the specified project and region
||Includes an instance into monitoring
||Retrieves the resources monitored by the Telemetry service
||Retrieves the metrics of the selected resources according to the specified parameters
||Removes an instance from monitoring
Further on this page, you can find the examples of the commands usage for Telemetry Service manipulation.
To start the Telemetry Service, run or2-manage-service (or2ms) command with the -a/--activate flag and the -s/--service-name parameter
with the "telemetry" value:
or2ms -p project -r region -s telemetry --activate
The command will launch a MEDIUM Ubuntu16.04 64-bit instance with Gnocchi and PostgreSQL as well as with a Gnocchi client (the Gnocchi shell utility allowing server manipulations via the
To deactivate the service, use the same command with the -d/--deactivate flag.
After the Telemetry server has been deployed, you need to add instances to it. The instances added to the monitoring will send their metrics to be stored in
the Telemetry server database.
To add an instance to Telemetry monitoring, use the following command:
or2starttel -p project -r region -i instance_id
The Telemetry server stores instance data referring to their original resource IDs. The original resource IDs are provided by the telemetry agents - the client
software installed on the monitored instance when it is added to Telemetry monitoring. To retrieve the telemetry agents and original resource IDs, use the
or2dtelag -p project -r region
In addition to the resource IDs, the command returns the date and time of instance inclusion into Telemetry monitoring::
If you specify the original resource ID in the or2dtelag command, the command will also return the list of metrics available for that agent:
or2dtelag -p project -r region -o original_resource_id
Using the original resource ID, you can retrieve the metrics from the Telemetry server with the following command:
or2tel -p project -r region -o original_resource_id -m metric_name -g granularity -a aggregation -d days -c count
The request must always specify the metrics name (the available metrics can be found in the response to the or2dtelag command).
Additionally, the command accepts the following parameters to filter the list of results to be returned:
||Measures granularity. Available values: minute, hour (default), day. For all granularity levels, the maximum of 1440 records can be returned
||Aggregation method. Available values: average, min, max, sum, count, std. Default: average
|-d, --days; -h, --hours
||Time period to return the metrics for. Cannot be combined with each other and the --from and --to parameters. Default: 1 day
|-f, --from; -t, --to
||Time period to return the metrics for. Cannot be combined with the --days and --hours parameters
If you are using the --from and --to parameters, make sure you enter the local time values. If you use GMT values, the returned data will be inaccurate.
Note that Telemetry as a Service returns metrics data in unnormalized form.
To stop monitoring of an instance, use the or2-stop-telemetry (or2stoptel) command:
or2stoptel -p project -r region -i instance_id
When an instance is removed from telemetry monitoring, no further data is collected from it. However, the previously collected data will remain in the database and
will be returned for the requests containing references to the removed instance and the corresponding time period.
The service Web UI is based on Grafana, an open-source platform for graphic data monitoring and analytics over time. Data from Gnocchi
is automatically transferred to Grafana and can be viewed on dynamic charts.
Grafana is installed on the same server which hosts Gnocchi during the activation of the Telemetry service. Use of Grafana involves no
additional virtual resources, thus having no influence on the service cost.
When Telemetry as a Service is activated, its information returned in response to the or2-describe-services (or2dser)
command contains the Web UI URL, login and password needed to access Grafana and view the instance metrics collected by Gnocchi.
When you log in to Grafana using the data from the or2dser command response, select New dashboard and configure it to display
data received from Gnocchi by setting the Panel Data Source field to gnocchi_database. The list of all connected resources
can be found under Resource ID and the metrics to be displayed on the chart ? under Metric name.
Each new query generates a line on the chart.:
The service usage price is defined by the price of the Telemetry Server VM.
The default parameters of a Telemetry Server VM are:
Therefore, the approximate monthly cost of a Telemetry Server usage in case of 100% and 24/7 load is about $42.2 in EPAM-BY2 region (as of 05/12/2017).
The price can vary depending on the region.
To get more detailed estimations, please, use our Cost Estimator tool.
More information on the Telemetry Service can be found in the EPAM Cloud Services Guide.
For detailed description of the Maestro CLI commands used to manage the Telemetry Service, refer to the
Maestro CLI User Guide.
The information on this and other analytics capabilities of EPAM Cloud Analytics is given on the
Cloud Analytics page.