A big thank-you to all our charm teams—Data, Observability, NoSQL, Telco, and many more—for joining our AMA sessions over the past cycle. Your questions and feedback are invaluable in shaping Ops, Pebble, Jubilant, and our broader tooling roadmap.
Key Issues from the AMAs
Testing & CI
Charmers wrestle with slow, flaky integration tests as well as unit test maintenance.
We’ve heard requests for more built‑in fixtures (all_idle
, @abort_on_fail
) and clearer guidance on migrating from Harness to ops.testing
.
Charm Libraries & Packaging
Dependency management (Pydantic v1 vs v2, nested libs), versioning workflows, and the PyPI vs fetch‑lib debate came up repeatedly.
Pebble & Observability
You’ve asked for better runtime metrics (OTLP log format, Pebble 1.19.0 health checks), dynamic cert reloads, and seamless log forwarding to Loki/OTLP collectors.
Relation Interfaces & Data Handling
Clarifying relation‑broken semantics (scale‑down vs removal) and richer helpers for inspecting event context remain high on your list.
Linking to Our Recent Work
Integration testing with Jubilant: We added the --keep-models
example and are tracking jubilant.all_idle and @abort_on_fail
helpers in issues #91 and #122.
Ops 2.x now bundles scenario tests under ops.testing, with improved examples and docs slated for the next doc‑revamp sprint.
Pebble 1.19.0 (Juju 3.6.4) delivers built‑in health checks and metrics; OTLP support is on our roadmap for the next cycle.
We’re drafting improved relation helpers in the Ops and gathering real‑world patterns to bolster documentation around broken vs departed events.
What’s Next?
We’ll be discussing which of these items to prioritise for the upcoming 25.10 cycle. Keep an eye on the Ops, Pebble, and Jubilant repositories for pull requests, and feel free to comment or submit issues if you have further suggestions.
Once again, thank you for your time, insights, and continued partnership in making the Juju ecosystem better for everyone.
— The Charm Tech Team