Observability and operability
Add the structured logs, spans, health checks, and alerts an incident needs to be readable.
The gaps only show up at 3am, when the person on call needs a log line that nobody added.
Someone has to add the trace ids, the structured logs, the health checks, and the spans that cross the boundary, and every new alert comes with its runbook.
Add health and readiness endpoints with tests
Endpoints respond, the unhealthy path is covered too
Attach correlation fields to one call path
Assertions that the fields survive every hop
Replace ad-hoc prints with the project logger
A rule blocks the old call, tests assert the new one
Instrument one endpoint group with metrics and spans
Span names asserted in tests, dashboard query included
Link logs, traces, and errors so one click moves between them
A single incident walked end to end across all three signals
Set a reliability target and alert before it is missed
Alert fires against a replayed window that breached it
Sort one boundary's errors into named categories
Each category carries a test and a stated remedy
Write the alert rule and the runbook it points at
Rule back-tested against the window it should have fired in
Put the agent team to work on your backlog
Connect a repository and assign SHIP your first issue. Your first missions are free.