Senior System Operations Engineer - 2 Headcount

Availability and Monitoring

Keep the products available and know before customers do when they are not. Own the monitoring and alerting in New Relic and UptimeRobot, tune it so an alert means something, and watch the trends that show a problem building before it becomes an outage.

• Availability against agreed targets, reported monthly

• Incidents detected by monitoring before a customer reports them

• Alerts that are acted on rather than ignored

Application Servers and the JVM

Run the Java application servers and the services around them. Heap and garbage collection settings, thread dumps, connection pools, session handling, and ActiveMQ and Elasticsearch where the products use them. Know how these fail as well as how they are configured.

• Server and JVM settings documented and held in source control

• Capacity problems seen in the trends before they cause an outage

• Restarts falling as their causes are removed

Database Operations

Look after the Oracle and SQL Server databases from the operations side. Backups that are restored by test, blocking and long running sessions caught and traced to their source, space and performance watched, and patching kept current, working with the developers on anything that needs a code fix.

• Backups restored by test on a schedule

• Blocking sessions traced to the code or job behind them

• Database patching current, with exceptions agreed

Releases, Maintenance and Change

Deploy releases and run planned maintenance through the agreed change process, inside the published windows, with a rollback that has been tested. Nothing is changed on a server without a record of what changed and who agreed it.

• Every change raised, reviewed and approved before it is made

• Releases deployed with a tested rollback

• Change failure rate tracked and falling

Incident Response and Root Cause

When something breaks, restore service first and communicate clearly while you do. Then work with the developers to find and remove the cause, and record it as a problem so the same incident does not come back.

• Time to restore service, tracked and falling

• Incidents you handled closed to root cause

• Recurrence rate on closed problems

Automation and Documentation

Treat manual steps as defects to be removed. Script repeat work in Bash, PowerShell or Python, use Ansible or equivalent for configuration, and keep runbooks current enough that someone else can follow them at 2am. Use AI tools to help write scripts and read logs, and review what they produce before it runs anywhere that matters.

• Repeat manual tasks automated, with the work they removed named

• Runbooks current and used by others

• Scripts held in source control, reviewed, and safe to run twice

Security, Access and Patching

Keep operating systems, Java runtimes and application servers patched, certificates renewed before they expire, and access granted on least privilege and reviewed rather than accumulated. Work to the shared infrastructure standards and with the platform team rather than in a local variation.

• Security patches applied within the agreed window

• No certificate expiry reaching production

• Access reviewed on a schedule, with leavers removed promptly

Requirements

•Solid experience running production Java applications on Linux and Windows, with personal accountability for uptime and incident response. This is the requirement we do not flex on.

•Java application servers and the JVM. Tomcat, JBoss or WildFly, heap and garbage collection tuning, thread dumps and connection pools. Experience of Intershop or another commerce platform is a strong advantage.

•Oracle and SQL Server from the operations side, covering backup and restore, sessions and locks, performance and patching.

•Automation and tooling. Bash, PowerShell or Python, Ansible or equivalent, CI/CD with Jenkins or Azure DevOps, Docker, and monitoring with New Relic or similar. ActiveMQ, Elasticsearch and Azure are beneficial.

•Hands-on use of AI tools in daily operations work, with a clear sense of where an automated change must not go unsupervised.