Move from container to job for cache bust

[adamrdrew]

More drama on the road to cache busting. We started with initContainer and didn't like it because the cache busted too long before app starting. Then we moved to another container in the pod, but we didn't like that because a container has to be long running and has to just loop forever, which is wasteful and confusing - as well as exposing the entire deployment to risk of failure for a bogus reason. So, this refactors AGAIN on a job.

[adamrdrew]

--- PASS: kuttl (421.10s)
    --- PASS: kuttl/harness (0.00s)
        --- PASS: kuttl/harness/basic-frontend (18.09s)
        --- PASS: kuttl/harness/networking (18.55s)
        --- PASS: kuttl/harness/whitelist (17.97s)
        --- PASS: kuttl/harness/storage (17.80s)
        --- PASS: kuttl/harness/ssl (134.10s)
        --- PASS: kuttl/harness/frontend-paths (17.83s)
        --- PASS: kuttl/harness/generate-nav-json (18.01s)
        --- PASS: kuttl/harness/bundles (19.92s)
        --- PASS: kuttl/harness/cachebust (153.71s)