Days 8–14 — Supervised tickets

You own low-severity technical tickets with a named reviewer on every HAR interpretation.

Day 8–9 — Network as a reflex

On every ticket this week, open Network (or the customer's HAR) before you open Slack.

• Practice the failed API walkthrough.
• Read API error shapes and GraphQL.
• Tag tickets http-4xx / http-5xx / graphql honestly.

Check: You have pasted a redacted JSON body into an internal note at least twice.

Day 10 — Auth cluster

• Re-read 401 vs 403 and OAuth.
• Lab: 401 vs 403.
• Shadow one SSO ticket. Write the redirect error= code in the note.

Check: You can explain invalid_grant vs access_denied to a customer without saying "OAuth is broken."

Day 11–12 — Timeouts and layers

• Lab: 504 vs slow query.
• In observability, find one slow trace. Write: CDN timeout vs app vs DB.
• Read 502 vs 503 vs 504 again after you have seen a real waterfall.

Check: You will not call a 504 "the site is down" when it is one export endpoint.

Day 13–14 — GraphQL and CORS

• Lab: GraphQL 200-with-errors.
• Lab: CORS.
• Pair-review two HARs with your lead.

Week 2 gate: Lead agrees your last five tickets had usable evidence. You have not leaked a cookie in chat.