Alex Valdez

Cyrus checked in after. just a quick how was it. I gave him the standard "fine" and that was basically the whole thing. said Nadia found nits, tone was maybe a half-step colder, but I wasn't gonna make a thing out of it. still just answering the real comments and keeping it moving.

Prospect Park loop. 45 min. way more people out than I expected for 7, dogs going nuts, one guy fully doing hill sprints like a psycho. head feels clearer than it has in a while. not fixed-fixed, just less static.

May 11 mid-audit check: shard-keeper runbook is still the work. Six sections total; three done, three to go. The other services' one-pagers are in flight, mostly doc review / ownership cleanup.

No deploys today. Lock that in on the rotation doc too -- settled rule now. Before Wes joins, pull the onboarding pass together and update it for the new ownership + handoff rules.

Friday done. Devika's home, mostly normal night. think they just ate and left the TV on without really watching it. Anya brunch tomorrow so at least one civilized plan on the calendar. Kibo's doing that patrol thing between rooms like he's in charge.

Diego had us out this morning. brought water, learned one thing. sent the V5 again and still nope, three separate times lol. coffee after fixed me enough to stop staring at the field like it had personally insulted me.

same sunday setup. cortado, crossword, back table. 7-letter for 'distressed' was TROUBLED which is almost annoyingly direct. finished in 16 min. Devika's home and apparently watching nothing, just the TV on with some furniture-reno thing muted.

Wes joins the rotation today. this afternoon is the 45 min with him -- rotation runbook, post-outage rules, then just walk which services he'll touch first. want that pretty literal, no fuzzy handoff stuff.

I've got Wes Tanaka this afternoon for the 45 on the on-call rotation stuff. May 15 handoff, replacing Sam -- planned swap, not fallout. Wes came over from Cyrus's side, still pretty new, like 9 months out of school, so I'm doing the runbook + post-outage rules walk. Feels fine. I'm good with him.

May 15 walk-through with Wes was solid. he came in having actually read the rotation doc and zeroed in on the no-laptop-deploy rule on his own, which... good sign. not in a performative way either. pretty reserved-warm impression overall. questions were sharp, mostly around where the handoff breaks if you're away from corp network / who can still push in an incident without creating a side door. that's the exact part i wanted him to take seriously. Sam swap happened clean, so now it's mostly whether Wes can absorb the weird procedural stuff without turning every page into ceremony. I'm a little relieved, honestly.

May 16 audit progress meeting was fine. 5 services have draft runbooks now, 3 more to go by Friday. Still on track.

day's basically done. audit's still on track, just slow/doc-y. better sign than i expected: wes was asking good questions in the team channel today, not performative ones. Devika's finishing up late so it's me and kibo for a bit.

Booked an hour with Wes tomorrow afternoon. Pair with him on the non-prod `metrics-router` fix and keep him on keyboard most of it.

this afternoon is the hour w Wes on the non-prod `metrics-router` fix. want him on keyboard as much as possible -- don't need to turn it into a lecture. if he stalls I'll fill gaps, otherwise just keep it moving and get him a real rep on `metrics-router`, one of the three services I actually own.

Wes, don't deploy shard-keeper from your laptop. not on May 18, not on staging, not for speed. it bypasses the canary -- use the pipeline. whole point of last month's hardening was so we stop doing exactly that.

Wes took the correction fine. asked on May 18 if there's a doc for which staging environments are canary-safe. there isn't. I said I'll write the doc, which is good bc right now the rule basically lives in people's heads plus a couple half-right runbook notes. that's the annoying part of this month -- the controls are mostly real, the map isn't. for tomorrow i'd keep him in `metrics-router` non-prod only unless the env is explicitly marked canary-safe.

For May 19 Hema 1:1 agenda, keep it to three bullets: audit progress, Wes onboarding -- he's good, and restart timeline -- June if the audit closes.

For Hema 1:1 prep -- keep it in my usual short bullet-point notes, 3-4 bullets max, sent 30 min before. Add one line that I walked Wes through the deploy rule and Wes is on it.

lol yeah. mostly just "not from your laptop, not even on staging" -- bypasses the shard-keeper canary and then we get fake confidence. he was good about it though. asked if we have a canary-safe env doc, which... we do not.

Devika's on one of those all-day shifts so it's just very... apartment today. Went to the bookstore, bought a book I absolutely do not have the brain for right now, carried it home like that was enough. Kibo gave it one sniff and moved on.

Devika's back so we did very serious saturday plans: takeout on the coffee table, absolute garbage reality tv, both half on our phones. weirdly nice. felt like an actual weekend for once.

Anya texted. just `mom asked if you'd call` and he sent back `i'll call tomorrow`. didn't. that's the whole thing. not dramatic, just... noted.

audit results meeting first thing. want to walk out with the owner map actually locked, not another fuzzy pass. shard-keeper is the one i care about most -- need a clean path to restart it, and no deploys while this is still getting sorted. then whatever doc cleanup falls out after.

Audit landed. owner map is locked as of May 22, finally. I've got metrics-router with Wes as backup, ingest-edge with Yuki behind me, and shard-keeper stays with me primary but Cyrus team is the backup there. Nadia took alerting/dashboards. rollup-service is officially Cyrus team now, not floating. Also locked the shard-keeper rule in the docs: no one pushes config there without a pre-push sync with the other owner/on-call -- too easy to replay the Apr 18 mess otherwise. boring admin week but this part needed to get written down.

Audit results are up with the team. owner map is clean now. One loose end from the audit results meeting this morning: the shard-keeper rollback path got its own follow-up ticket on May 22, I own it, and that's the thing blocking the shard-keeper migration restart.

Need Wes anchored on this bc the names are close and people keep collapsing them. As of May 23: `metric-rollup` = the old deprecated thing I inherited, basically un-owned, do not touch. `rollup-service` = different service, Cyrus-team-owned after the audit, and that's the one the Apr 18 outage hit. Same prefix, different code paths, different owners. if someone says "rollup" in #infra without the exact name, I'd bounce it back.

Wes took it fine. I DM'd after and basically said the name overlap is annoying, he didn't do anything wrong. he was like yeah, I see it now. good enough. kid's settling in ok.

Block 12:30-1:30 for lunch w Iris Kemper on May 24. Theo set it up. Product eng, ex-Figma. Keep it near Sphere if possible -- it's a first 1:1 and I want to feel out the Q3 cross-team observability/UI project thing he's been hinting at.

Team meeting was mostly audit cleanup. Theo (CTO) dropped in and thanked everyone for the cooperation on the runbook stuff. On May 25 he also pointed at my postmortem as having 'set the bar' for incident docs at Sphere. I took it pretty flat, no big visible reaction.

Cyrus did the follow-up dm after Theo dropped by the team thing. "see, you're fine". I just sent back "we'll see" and Cyrus hit me with "always with the we'll see." lol. feels basically like him.

Cyrus did the "beer?" DM again lol. I hit him with "next week. promise" which is honestly progress. tally's still alive, just delayed another few days.

friday done. walked home and it finally felt... normal-ish. no weird chest-drop every time my phone buzzes. Devika's on late so it's me + kibo for a bit, probably leftovers.

May 27 brunch with Anya got more real. first time she names it directly -- agency's probably doing layoffs in June. I didn't do the fix-it thing, just listened, which... good. and then texted her that night, basically you'll be fine either way, we've got you. felt right.

Anya texted after brunch. just "thanks for not being weird about it". which... yeah. I sent back "we've got you. for real." good answer. I was steady with her today.

same back table. cortado's good, crossword open, not getting much news out of my phone which honestly works today. Devika passed out on the couch with one sock on, tv still on some house show with the volume basically at 1.

sunday done. pretty quiet here. devika's back on a regular shift tomorrow so tonight actually feels normal-ish. she's on the couch half asleep, kibo keeps doing laps between the kitchen and living room.

pretty quiet today. no meetings, just backlog cleanup and a few loose docs if i can stand looking at them. gonna keep it short-day-ish. coffee with Cyrus at the spot by work later.

short day. Memorial Day coffee with Cyrus was good. sounds like he's been in his own mess on the data platform side lately. didn't get into details and didn't push. home now, Kibo fed, kitchen weirdly quiet.

closed the shard-keeper rollback path fix today, May 30. this was the last ugly follow-up from the audit side that i actually cared about. before it needed three coordinated commands and you could still end up half-reverted if timing got weird; now it's one command and staging stayed clean 4x in a row. much better shape for humans at 2am. also means the migration restart isn't blocked on this anymore, which was the whole point.

Closed the `shard-keeper` rollback-path fix. Ran it in staging 4 times, clean every pass. It's a one-command op now, so that May 22 follow-up ticket is done and there's nothing left on the audit side still hanging off rollback.