0.1.2

Bug Fixes

• kubernetes: recover from transient watch errors when waiting for the analysis pod. The pod watch was pinned to the resource version from the pod's Create call, which the apiserver can expire (HTTP 410 Gone) by the time the watch is serviced. This surfaced as a fatal watch error and failed otherwise-healthy analyses. The runner now re-establishes the watch from a fresh resource version on transient errors instead of failing. (24807e6)

Full Changelog: 0.1.1...0.1.2