Tracing team - Pulse #13 of 2026
The Team
We are part of the Observability team, alongside Core and Service Mesh.
The Work
Preparing for our release
feat: Added Juju secrets support for sensitive Grafana custom config fields: grafana-k8s
feat: Rolled out the charms blueprint across the observability portfolio
feat: Improved the blueprint changelog with track handling: observability
fix: Stopped continuous databag churn on the grafana-dashboard relation by only updating when templates change: opentelemetry-collector, opentelemetry-collector, opentelemetry-collector-k8s, opentelemetry-collector-k8s, grafana-k8s
fix: Sorted nginx server addresses to prevent needless coordinator reconfiguration: mimir, loki, tempo, cos-coordinated-workers
fix: Sorted peer names in prometheus_remote_write to eliminate duplicated alerts, and bumped the library in OTel Collector: prometheus-k8s, opentelemetry-collector, opentelemetry-collector-k8s, opentelemetry-collector-k8s, opentelemetry-collector
fix: Resolved constant Alertmanager service restarts: alertmanager-k8s, alertmanager-k8s
fix: Repaired mimirtool TLS failures, stale hash on exec errors, and missing leader guard, and updated write endpoints: mimir, mimir, mimir
fix: Coordinators no longer reconcile on pebble-check-failed/recovered events: cos-coordinated-workers
fix: Mimir coordinator now trusts Mimir’s CA when syncing rules over TLS: mimir, mimir
fix: Loki ruler API accepts syncing alerts (backport): loki
fix: OTel Collector snap now applies path_exclude to log jobs: opentelemetry-collector, opentelemetry-collector
fix: Excluded juju_charm from alert rule expressions to avoid false labelling: loki-k8s
fix: Charm tracing now uses the public Python API for version compatibility: tempo
docs: Clarified that Loki ingestion-rate-mb is shared across distributors: loki
chore: On-demand release workflow triggers and wokeignore cleanup for Tempo, plus an update workflow for apache-exporter-snap: tempo, tempo, tempo, apache-exporter-snap
chore: Upgraded grafana_source library to v1 for stable datasource UIDs across coordinators and charms: prometheus-k8s, parca-k8s, pyroscope, loki-k8s, alertmanager-k8s, tempo, loki, mimir
Others: we continued our secondments with other Canonical teams, and kept improving our internal Solutions QA tooling.
Feedback welcome
As always, feedback is very welcome! Feel free to let us know your thoughts, questions, or suggestions either here or on our Matrix channel.
That’s all for this time! See you again in two weeks!