01
- Rescue
- mTLS blocking traffic, sidecar injection failing, or a mesh upgrade that left two revisions fighting. We stabilise and explain second.
02
- mTLS rollout
- Permissive mode first with telemetry to find the plain HTTP callers, policies written against real data, then STRICT turned on without losing requests in the switch.
03
- Traffic management
- VirtualService, DestinationRule and Gateway written for canary releases, blue green deploys, mirroring and retries that are safe rather than aggressive.
04
- Ambient migration
- Moving from sidecar mode to ambient where it fits, with capacity planning, telemetry parity and a rollback path at each step.
05
- Observability
- Kiali, Prometheus and distributed traces wired together so a 503 in the mesh tells you where it came from, not just that it happened.
06
- Managed Tech
- A standing arrangement covering Istio minor upgrades, Envoy CVE response, policy review and a mesh engineer available when a routing change at 11pm takes production down.