You can view your service metrics from Timescale Cloud's metrics dashboard. This dashboard gives you service-level information, such as CPU, memory, and storage usage.
You can also view your query-level statistics by using the pre-installed
pg_stat_statements extension from a PostgreSQL client.
Timescale Cloud provides a Metrics dashboard for managing your services. You can
see the Metrics dashboard in your Timescale Cloud account by navigating to the
Services section, clicking the service you want to explore, and selecting the
You can view metrics for your services for any of these time ranges:
- Last hour, with one minute granularity
- Last 24 hours, with one minute granularity
- Last seven days, with one hour granularity
- Last 30 days, with one hour granularity
To change the view, select the time range from the drop-down menu.
Additionally, you can turn automatic metric refreshes on and off. When automatic metric refresh is on, the dashboard updates every thirty seconds.
In some cases, gray vertical bars display on the metrics dashboard, like this:
This indicates that metrics have not been collected for the period shown. It does not mean that your Timescale Cloud service was down.
Timescale Cloud continuously monitors the health and resource consumption of all
database services. You can check your health data by navigating to the
tab in your service dashboard. These metrics are also monitored by the Timescale
If your database exceeds a storage threshold of available resources, some automated actions are triggered, including notifications and preventative actions.
When your disk usage exceeds certain thresholds, you receive an email notification. These notifications occur at:
So that you aren't overwhelmed by automated messages, the alerting thresholds use low and high watermarks, and we limit the frequency of messages we send you about a particular service.
If your database continues to increase in size past these thresholds, automated overload protection is activated when your disk becomes 99% full. When this happens, your database is put into read-only mode, you receive an email notification, and the Timescale Cloud console shows the changed status.
When your disk is in read-only mode, you can still query your database, but you cannot add any new data to it. This ensures that your disk does not fill up to 100%, and thus prevents the database from crashing due to an out of memory (OOM) error.
With your database in read-only mode, you need to decide if you are going to increase your storage capacity, or reduce the size of your database. When you have done that, you can also add a retention policy, or turn on compression, to avoid the problem occurring again in the future.
pg_stat_statements extension gives you query-level statistics for your SQL
statements. It comes pre-installed with Timescale Cloud.
For more information about
pg_stat_statements, see the PostgreSQL documentation.
You cannot currently enable
track_io_timing for your database. Statistics that depend on
track_io_timing, such as
blk_write_time, are not collected.
You can view statistics for your queries through the
extension, which provides a
pg_stat_statements view. The recorded statistics
include the time spent planning and executing each query; the number of blocks
hit, read, and written; and more.
You can query the
pg_stat_statements view as you would any PostgreSQL view.
The full view includes superuser queries, which are used by Timescale Cloud to
manage your service in the background. To view only your
queries, filter by the current user.
Connect to your database using a PostgreSQL client, such as
SELECT * FROM pg_stat_statements WHERE pg_get_userbyid(userid) = current_user;
pg_stat_statements, you can view performance statistics that help you
monitor and optimize your queries.
Here are some sample scenarios to try.
Identify the 5 longest-running queries by their mean execution time:
SELECT calls,mean_exec_time,queryFROM pg_stat_statementsWHERE pg_get_userbyid(userid) = current_userORDER BY mean_exec_time DESCLIMIT 5;
Identifying queries with highly variable execution time:
The relative standard deviation, or the standard deviation expressed as a percentage of the mean, measures how variable the execution time is. The higher the relative standard deviation, the more variable the query execution time.
SELECT calls,stddev_exec_time/mean_exec_time*100 AS rel_std_dev,queryFROM pg_stat_statementsWHERE pg_get_userbyid(userid) = current_userORDER BY rel_std_dev DESCLIMIT 5;
For more examples and detailed explanations, see the blog post on identifying
performance bottlenecks with
Found an issue on this page?Report an issue!