[feature] Add archive specific metrics

[spali]

They are basically free, as already provided by the actual command results.
I find them actually much more interesting than the current sizes of the whole repo... as they allow to conclude to some potential problems or trends...
i.e. :

• compare archive file counts.... if they differ extrem ... someone deleted everything by accident or by malicious code.
• allow to track data increase, compression and deduplication a bit different. Per archive can give different insights.
• duration changes can also show potential problems.

for sure, this metrics only makes sense if you have one repo per backup and the archives represents snapshots of the same data.... but this was anyway a assumption for using last archive data per repo i.e. in borg_last_backup_timestamp.

Sorry for not providing the actual dashboard to this metrics, as I didn't had time to make them proper for the PR.

[maxim-mityutko]

Those are indeed interesting use cases!
Please add tests:

borgmatic-exporter/tests/test_metrics.py

Lines 44 to 70 in 0d8a95c

	def test_registry(self, registry):
	result = metrics.create_metrics(registry=registry)._names_to_collectors
	assert "borg_total_backups" in result
	assert "borg_total_chunks" in result
	assert "borg_total_compressed_size" in result
	assert "borg_total_size" in result
	assert "borg_total_deduplicated_compressed_size" in result
	assert "borg_total_deduplicated_size" in result
	assert "borg_last_backup_timestamp" in result
	@pytest.mark.parametrize(
	"metric, repo, expect",
	[
	("borg_total_backups", "/borg/backup-1", 2.0),
	("borg_total_backups", "/borg/backup-3", 0.0),
	("borg_total_chunks", "/borg/backup-1", 3505.0),
	("borg_total_compressed_size", "/borg/backup-1", 3965903861.0),
	("borg_total_size", "/borg/backup-1", 8446787072.0),
	("borg_total_deduplicated_compressed_size", "/borg/backup-1", 537932015.0),
	("borg_total_deduplicated_size", "/borg/backup-1", 1296544339.0),
	("borg_total_deduplicated_size", "/borg/backup-2", 21296544339.0),
	],
	)
	def test_individual_metrics(self, collect, metric, repo, expect):
	registry = collect
	actual = registry.get_sample_value(name=metric, labels={"repository": repo})
	assert actual == expect

Would be good to see how this can be visualized as well, but this can also be done in the next release.
Again, thanks for contributing!

[spali]

missed the tests again... this time I checked at least the coverage but that didn't change and the above tests don't check actual code lines.

[spali]

not sure how you choosed the examples for the parameterized test... just added all new metrics for backup-1 and backup-2.
Hope you are ok with that.

[maxim-mityutko]

not sure how you choosed the examples for the parameterized test... just added all new metrics for backup-1 and backup-2. Hope you are ok with that.

The values are pretty much random.

Please black the files, otherwise the checks are failing.
https://github.com/maxim-mityutko/borgmatic-exporter/actions/runs/11263638182/job/31323139133?pr=22

[maxim-mityutko]

Thanks! Will release this shortly.

[maxim-mityutko]

@spali I have releases the v0.2.7, do you by any chance have the updated Grafana dashboard that exposes these metrics?

[spali]

No, didn't had time to prep some dashboards. But they may be also very individual, especially for alerting.

[maxim-mityutko]

Makes sense.
I've updated the existing dashboard though to include new metrics and changed the structure a little bit.