Databricks
Security / Reliability EngineerAppliedAug 6, 2026

Contain outbound connection exhaustion in peak tasks

Restricted connectivity can fail from volume even when every destination is approved.

The challenge is to preserve a controlled boundary while preventing routine workload behavior from overwhelming it.

We’re not trying to reach anything new—we just can’t finish the load before the handoff.

Cheruiyot Seyoum · Lead Data Engineer

Owns high-volume batch transformations that load curated results into an approved operational database.

What pulls against what

  • connection capacity vs. egress control
  • throughput vs. retry amplification
  • fast vendor relief vs. root-cause clarity
  • shared limits vs. tenant urgency

What is at stake

Peak batch workloads are failing at a shared network boundary. The fix must improve completion without loosening egress controls

Why Databricks

At Databricks, this can matter when managed compute must reach customer-approved services predictably.

Written for

Reliability engineerNetwork security engineerDistributed systems troubleshooter

This is the setup. The work is inside.

Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.