Guido Trotter

1.6: term.color(<number>) fails as a function if terminal doesn't do styling

1

When in non styling mode one can use term.color(i) + "string" + term.normal but not term.color(i)("string") Thanks! Guido

This adds a dependency on python-dateutil to calculate years and months lengths, depending on the current date. Alternatively we can just go for a fixed arbitrary time length for years...

test_get_interfaces contains duplicate code from the superclass. Simplify it by just adding the local mock wrapper but without copying the code.

I filed this as https://github.com/sensu/core/issues/27 and it has an associated pull request https://github.com/sensu/core/pull/24 which is really a one-liner plus a protobuf recompile. ## Feature Suggestion The idea is to keep...

If there are a lot of silences Silences.query, for queries without q.ids, first copies all their pointers to a slice, causing a lot of allocations, and then does it again...

This summarizes the principles for HA as stated by Ben, an old blog post by Julius explaining how it works[1], and the current implementation. Signed-off-by: Guido Trotter [1] https://promlabs.com/blog/2023/08/31/high-availability-for-prometheus-and-alertmanager-an-overview/#ha-for-the-alertmanager

Calling QIDs with an empty list of IDs is meaningless, and the result may be unexpected (at the moment it would return all silences, but one might expect it to...

loadSnapshot: update matcher index properly while not holding lock

3

Since we're building the state separately, we will do the same with the matcher index, and then update it into the Silences with the lock held.

Improve silence benchmarks

1

- Add some benchmarks for GC - Add benchmarks that call Mutes while adding some silences, to better test how it would be called in the real world - Change...

Currently if errors are found GC aborts, but will keep bumping into the same errors, forever. With this change we count the errors, but also remove the problematic entries, thus...