DevOps Engineer
Frequently asked questions about hiring for the DevOps Engineer role, plus the mistakes that most often derail it.
Common hiring mistakes for this role
DevOps roles are particularly often mis-scoped at German SMBs. The following anti-patterns regularly cost 6 to 12 months of onboarding and lead to early resignation.
Advertising DevOps as a pure SysAdmin role
A job posting with Linux server administration, SSH, backups, patch management as the main responsibilities filters out modern DevOps profiles immediately. You attract profiles that cannot work autonomously in cloud-native setups, and you deter profiles with Kubernetes, IaC and observability depth. Instead, describe the real task mix: platform engineering, pipeline ownership, observability, incident response, enablement of the app teams. If the role really is predominantly SysAdmin, advertise it as a System Engineer, not as DevOps.
Hiring on corporate pedigree instead of production ownership
A top graduate of a renowned university with 2 years at a DAX corporation or a big-tech branch is not automatically more productive than a self-taught engineer with 5 years of on-call at a startup or SMB. Large structures give their platform profiles clear contracts between teams, dedicated SRE functions and solid internal tooling; at an SMB this scaffolding is usually missing, and the lack of production autonomy becomes a burden. Weight the incident-walkthrough interview and the reference check on autonomy more heavily than the pedigree on the CV.
Demanding multi-day take-home assignments
A take-home assignment of 8 or more hours actually consumes 24 hours (with emotional investment), demotivates the best profiles (who have other options in parallel) and delivers no better signal than a well-run incident walkthrough of 45 min plus a 30 min architecture discussion. You want to measure the quality of the reasoning, not the completeness. If a practical exercise still seems necessary, cap it at a maximum of 2 hours with clear scoping and accept incomplete but well-reasoned solutions.
Making the on-call expectation transparent only after the contract is signed
A production-owning DevOps role with no clear statement on on-call, standby model, standby pay, average pager load and escalation paths in the posting leads to early resignation within 6 to 12 months. State the model openly: standby window (e.g. 1 week every 6 weeks), pay (standby allowance plus call-out hours), realistic pager frequency, escalation option. Profiles that reject on-call on principle are often out of place in an SMB DevOps role; better to filter them out now than to lose them after 6 months.