Automated check

c2cgeoportal applications include functionality for testing and assessing that the application is functioning correctly, i.e. that its web services respond as expected.

For this purpose, c2cgeoportal provides a checker and a check_collector. These (especially the collector) are meant to be used by a monitoring system like Nagios to check that the application is alive.

They integrate into c2cwsgiutils’s health check service. The return codes are:

• 200 => OK

• 500 => Error

Note

Check collector can only check pages that are on the same server as itself.

You can access the health_check service with this kind of url:

<project_url>/c2c/health_check?max_level=3

The levels used in the default configuration (recommended settings):

• 1: Fatal checkers (main database connection, alembic versions).

• 2: Vital checkers (other database connection, redis connexions, versions).

• 3: Quick internal checks.

• 4: Slow service checks (phantomjs checks, theme).

• 5: Service checks (caution: these can result in costs, for example, print a page).

• 10: collector of different sites.

See also: https://github.com/camptocamp/c2cwsgiutils/#health-checks

The recommended URL that should be checked by services such as Pingdom or StatusCake:

<project_url>/c2c/health_check?checks=check_collector

The recommended URL to use to validate a migration (checker_url in project.yaml):

<project_url>/c2c/health_check?max_level=9

Checker

The checker uses the following configuration structure in vars.yaml:

vars:
    # Global configuration
    ...
    checker:
        # Checker configurations

Note

If some of the checked services require an authenticated user, the forward_headers configuration allows to forward Cookie or Authorisation headers from the underlying requests.

checker:
    forward_headers: ['Cookie', 'Authorisation']

Note

The checker assumes that it can access the c2cgeoportal services via http://localhost. If this is not allowed on your server, you can override this behavior as follows. In your vars file, add the following:

vars:
    checker:
        rewrite_as_http_localhost: False

Now, in your configuration file project.yaml, instead of defining the checker_path, define a checker_url with the full URL to be used, for example:

...
host: ${host}
checker_url: https://${host}${entrypoint}check_collector?
...

print

Check the print version 3.x (try to print a page).

Uses as spec the spec from configuration.

fulltextsearch

Check that the Full-text search service returns an element.

Uses the search from configuration as text to search.

themes

Check that the theme has no error for all interfaces present in the database.

Uses the themes configuration:

themes:
    params:
        # Dictionary that represents the query string
    <interface>:
        params:
            # Dictionary that represents the query string
    level: 2

lang

Check that all the language files are present.

Uses the global configuration available_locale_names, and the checker configuration files, an array of strings that must be in [ngeo].

routes

Check some routes, configured in routes as array of objects with:

• name: the route name.

• display_name: the name to be displayed.

• params: the used query string as a dictionary.

• level

In the configuration, we can also fill the routes_disable to disable some routes.

phantomjs

Check with phantomjs that the pages load correctly without errors, using the routes configuration as an array of route names to check:

• name: the route name.

• params: the query string to use, configured as a dictionary.

• level

Infrastructure

If you experience connection issues with your checker, the following configuration options may be useful for you:

vars:
  checker:
    [base_internal_url]: URL like http://localhost:8080
    [forward_host]: <True|False>

Check collector

Used to collect checks from a different instance in the parent/children structure. It is useful to perform a set of checks all at once.

The check collector uses the following configuration structure:

vars:
    # Global configuration
    ...
    check_collector:
        max_level: 1
        level: 2
        hosts: []

The max_level is the default max_level parameter used for every host. The max_level option can be set for a host to override it.

To add a host:

vars:
    check_collector:
        hosts:
          - display: Child
            url: https://{host}/child
            max_level: 1

update_paths:
  - check_collector.hosts