.. _using_managed_services:

Monitoring managed cloud databases
==================================

Although all cloud service providers offer some kind of built-in instrumentation and graphs, they're mostly rather conservative
in this as not to overconsume server resources and not to overflow and confuse beginners with too much information. So
for advanced troubleshooting it might make sense to gather some additional metrics on your own, especially given that
you can also easily add custom business metrics to pgwatch2 using plain SQL, for example to track the amount of incoming
sales orders. Also with pgwatch2 / Grafana you have more freedom on the visual representation side and access to around
30 prebuilt dashboards and a lot of freedom creating custom alerting rules.

The common denominator for all managed cloud services is that they remove / disallow dangerous or potentially dangerous
functionalities like file system access and untrusted PL-languages like Python - so you'll lose a small amount of metrics
and "helper functions" compared to a standard on-site setup described in the :ref:`previous chapter <preparing_databases>`.
This also means that you will get some errors displayed on some preset dashboards like "System Stats", "DB overview" and thus
will be better off using "lighter" dashboards like "DB overview Unprivileged" tailored specially for such a use case.

pgwatch2 has been tested to work with the following managed database services:

Google Cloud SQL for PostgreSQL
-------------------------------

* No Python / OS helpers possible. OS metrics can be integrated in Grafana though using the `Stackdriver <https://grafana.com/docs/grafana/latest/datasources/google-cloud-monitoring/>`_ data source

* "pg_monitor" system role available

* pgwatch2 default preset name: "gce"

* Documentation: https://cloud.google.com/sql/docs/postgres

NB! To get most out pgwatch2 on GCE you need some additional clicks in the GUI / Cloud Console "Flags" section to enable
some common PostgreSQL monitoring parameters like *track_io_timing* and *track_functions*.


Amazon RDS for PostgreSQL
-------------------------

* No Python / OS helpers possible. OS metrics can be integrated in Grafana though using the `CloudWatch <https://grafana.com/docs/grafana/latest/datasources/cloudwatch/>`_ data source

* "pg_monitor" system role available

* pgwatch2 default preset names: "rds", "aurora"

* Documentation:

    https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/CHAP_PostgreSQL.html
    https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/Aurora.AuroraPostgreSQL.html

NB! Note that the AWS Aurora PostgreSQL-compatible engine is missing some additional metrics compared to normal RDS.


Azure Database for PostgreSQL
-----------------------------

* No Python / OS helpers possible. OS metrics can be integrated in Grafana though using the `Azure Monitor <https://grafana.com/docs/grafana/latest/datasources/azuremonitor/>`_ data source

* "pg_monitor" system role available

* pgwatch2 default preset name: "azure"

* Documentation: https://docs.microsoft.com/en-us/azure/postgresql/

NB! Surprisingly on Azure some file access functions are whitelisted, thus one can for example use the "wal_size" metrics.

NB2! By default Azure has **pg_stat_statements** not fully activated by default, so you need to enable it manually or via
the API. Documentation link `here <https://docs.microsoft.com/en-us/azure/postgresql/howto-optimize-query-stats-collection>`_.


Aiven for PostgreSQL
--------------------

The `Aiven developer documentation <https://developer.aiven.io/docs/products/postgresql/howto/monitor-with-pgwatch2.html>`_ contains information on how to monitor PostgreSQL instances running on the Aiven platform with pgwatch2.