This extends the check so it can hold any arbitrary data. That is then
used in the generate report command, that finds all checks tagged with
report tag, executes them and gather their data for the final YAML
creation.

This commit also includes two examples of metrics that uses database as
the Data source.

With this new child class, report checks can share helpers easily. Also
the parent Check class, which is used elsewhere, does not need to be
extended by the data storage.

Earlier, all Reports were also considered Checks, so they showed up in
health checks. This change now fully isolates Reports and Checks while
all the code is shared.

Add recurring logics report

Add grouping report

no idea why this was ever there

we do not deploy dynflowd, only dynflow_sidekiq since
https://projects.theforeman.org/issues/28421 (RPM)
https://projects.theforeman.org/issues/28068 (DEB)

…Metadata command

We already excluded them during backup, so there is nothing to exclude
during restore and we do not support restoring from backups of super old
versions that did not exlude things.

See theforeman/actions#50

Fixes: 7c366e4

This marker file was removed in
theforeman/puppet-candlepin@18bca42
and Candlepin now manages its migrations on its own.

The installer will do that for us already, no need to duplicate work.

1/ it's not present in most cases
2/ even if it was, puppet/installer would manage it just fine

This was introduced in 8fda535 and
later changed in e6547a9 but it never
has matched a file as the symlinks are
`dynflow_sidekiq@<instance>.service` while the code only looked for
`<instance>.service`

Given nobody complained about dynflow not working after a restore in the
last 4 years, I am rather sure we can just not back those up and still
be happy.

We introduced this metadata field in
theforeman#536 which landed in
0.9.1, which is old enough to rely on and not care for backups that
don't have this metadata field

1/ reloading of systemd should come before starting services
2/ we don't backup any systemd files, so a reload is not required
3/ the installer will start all required services for us

when we were in the pulp 2to3 migration, users could end up with backup
that had both pulp2 and pulp3 databases, but only really had to restore
pulp2. now that pulp2 is gone, we can always assume that if a backup has
pulpcore DB, the user wants it restored

Fixes: 9ac8a20

Fixes: e3c802b

We don't enable maintenance mode anymore for backups, so we don't need
to disable it.

Fixes: f0e78f7

all others do not need the postgres user to be able to read the backup
for restore

…ion'"

    [root@centos8-stream-katello-nightly ~]# foreman-maintain update check
    Checking for new version of rubygem-foreman_maintain...
    Nothing to update, can't find new version of rubygem-foreman_maintain.
    undefined method `current_minor_version' for nil:NilClass

    I, [2024-06-25 11:46:51+0000 #12497]  INFO -- : Running foreman-maintain command with arguments [["update", "check"]]
    E, [2024-06-25 11:46:54+0000 #12497] ERROR -- : undefined method `current_minor_version' for nil:NilClass (NoMethodError)
    /usr/share/gems/gems/foreman_maintain-1.6.10/definitions/scenarios/satellite_upgrade.rb:7:in `block (2 levels) in upgrade_metadata'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/concerns/metadata.rb:230:in `instance_exec'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/concerns/metadata.rb:230:in `block in evaluate_confines'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/concerns/metadata.rb:229:in `all?'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/concerns/metadata.rb:229:in `evaluate_confines'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/concerns/metadata.rb:204:in `present?'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/detector.rb:62:in `block in available_scenarios'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/detector.rb:61:in `select'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/detector.rb:61:in `available_scenarios'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/concerns/finders.rb:29:in `find_scenarios'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/update_runner.rb:25:in `find_scenario'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/update_runner.rb:74:in `run_phase'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/cli/update_command.rb:30:in `execute'
    /usr/share/gems/gems/clamp-1.3.2/lib/clamp/command.rb:66:in `run'
    /usr/share/gems/gems/clamp-1.3.2/lib/clamp/subcommand/execution.rb:18:in `execute'
    /usr/share/gems/gems/clamp-1.3.2/lib/clamp/command.rb:66:in `run'
    /usr/share/gems/gems/clamp-1.3.2/lib/clamp/subcommand/execution.rb:18:in `execute'
    /usr/share/gems/gems/clamp-1.3.2/lib/clamp/command.rb:66:in `run'
    /usr/share/gems/gems/foreman_maintain-1.6.10/lib/foreman_maintain/cli.rb:40:in `run'
    /usr/share/gems/gems/clamp-1.3.2/lib/clamp/command.rb:140:in `run'
    /usr/share/gems/gems/foreman_maintain-1.6.10/bin/foreman-maintain:15:in `<top (required)>'
    /bin/foreman-maintain:23:in `load'
    /bin/foreman-maintain:23:in `<main>'

Fixes: 1ae71c3

Fixes: 041af1e

Heredoc has been supported since Ruby 2.3 and the gemspec already
mandates Ruby 2.7+.

Not all output supports colors. For example, cron doesn't. This copies
the autodetection from Kafo to disable colors if needed.

when restoring incremental backups, you can run in the situation where
the on-disk representation of the database changes between the restore
of the base and the increment, leading to a broken database.

by using logical DB dumps, we ensure that the restore always work at the
cost that increments are slightly bigger as they now always include a
full DB dump

this also allows simplification of the code as we now only have one type
of DB backup to support

optionally allow waiting for them to complete

this ensures more consistent on-disk data

In 5c2a705 the logic was broken from
"managed cron and not (downstream older than 6.3)" to "managed cron and downstream"

Fixes: 5c2a705

the procedures are already confined via `for_feature :cron`

the feature is already confined to that option being true