network-engineer — someone who gets blamed for everything

network-engineer [--on-call] [--tired] [--blamed] [--nocredit] ...

[ticket...] [vendor-call...] [blame-session...]

A network-engineer is a highly-skilled professional responsible for designing, implementing, and maintaining the infrastructure that allows all other engineers to do their jobs — and then receive credit for it.

The network-engineer exists in a state of perpetual blame. Traffic moves: nobody notices. Traffic stops: the network-engineer is paged within 90 seconds, regardless of root cause.

The primary output of a network-engineer is uptime. Uptime is invisible. Downtime is very visible. This asymmetry is a load-bearing feature of the role.

--on-call Accept pages at 3am for issues that are not networking

--troubleshoot Begin from layer 1 and work upward, as god intended

--ping Verify connectivity (will be blamed for ICMP being blocked)

--traceroute Trace the path (30 hops, no route to career-fulfillment.io)

--bgp Open a BGP session (may trigger global routing incident)

--ospf-area 0 All roads lead through area 0, including your problems

--mtu 1492 Discover this was the issue all along

--blame-dns Invoke DNS as the root cause (correct ~40% of the time)

--no-credit Default. Cannot be disabled.

0 Outage resolved, root cause identified

1 Outage resolved, root cause unknown ('it came back on its own')

2 Promoted to management (see: SANITY DEPLETED)

130 Ctrl-C during a maintenance window that ran long

255 Fiber cut by construction that was in the CAD drawings

TICKET_PRIORITY Always P1, regardless of actual impact

VENDOR_SUPPORT_ETA 2-4 business hours (add 72 hours for accuracy)

CHANGE_WINDOW Scheduled 6 weeks out; incidents don't check calendars

ROOT_CAUSE DNS (see also: BGP, MTU, the patch cable)

CREDIT_RECEIVED unset

The user described in DESCRIPTION often receives the following:

- 'Can you just make the internet faster?'

- 'Why can't I just use my home WiFi at the office?'

- 'What do you mean the problem is the patch cable?'

- 'Is it possible the network is too secure?'

These are known issues with no planned fix.

sre(1), devops(1), sysadmin(1), help-desk(1), blame(1), traceroute(8),

bgp(8), ospf(8), dns(8), mtu(4), patch-cable(3), change-management(7)

The network-engineer utility first appeared in the 1970s.

It has been indispensable ever since.

It has never been properly thanked.

Everyone who has ever stared at a Wireshark capture at 4am.

Everyone who has ever answered 'have you tried turning it off and on again' seriously.

Everyone who has closed a ticket with 'no issue found — self-resolved.'