18 YOE in IT (5.5 as Observability Engineer, AKS/New Relic) trying to formalize the jump to SRE — what actually matters in interviews?
18 years in IT overall (started in helpdesk/lab admin, 10 of those years at Juniper Networks across QA/test engineering), the last 5.5 as an Observability Engineer on a SaaS platform running on AKS. Day to day is mostly New Relic — alert design, dashboards, APM, some NRQL work that goes deeper than the defaults — plus Fluent Bit for log shipping and Python/PowerShell for internal tooling and custom metrics pipelines.
My contract winds down at the end of this year, so a transition that used to be a "someday" goal is now an active, time-boxed one. I want to move into a proper, production/customer-facing SRE role rather than just another observability/monitoring title, and I'd rather close real gaps now than find out about them in an interview.
Some of what I've actually owned: alert frameworks built around FACET-based NRQL (steady-state dashboards faceted by container, not pod, learned that one the hard way), a New Relic region migration, RCA work using distributed tracing to find gaps between synthetic and APM signal, and building custom metrics pipelines that feed New Relic from SQL/PowerShell.
Where I'm less sure of myself: hands-on K8s admin depth vs. "I can read a dashboard and explain a CrashLoopBackOff," real infra-as-code (Terraform/ARM) vs. just monitoring infra someone else provisioned, and owning SLOs/error budgets rather than just building the dashboards that report on them.
For people who made a similar observability → SRE jump:
- What was the actual gap that mattered in interviews — not the resume gap, the real one?
- Is CKA worth the time investment, or do interviewers not really probe that deep on K8s admin for a production SRE role?
- How much IaC depth do you actually need to do the job vs. just be able to speak credibly to it?
Appreciate any honest input, especially from people who've sat on the hiring side of this transition.