A launch-day email monitoring checklist is two resolvers, a foreign probe, history, a leftover MX check, and a morning re-query—written per zone. It is not a self-send. It is not “we’ll watch Gmail.” It is not ten cuts on one Friday without sheets. Exclusive MX. Unique subjects. HOLD review if unknowns are on. Filename every screenshot with the zone and the date.
Quick answer for email migration monitoring checklist
Print this page onto the ticket. Product path: delivery recovery, troubleshooting, docs, features, migration planner. IETF RFC 5321 — Simple Mail Transfer Protocol. IETF RFC 1035 — Domain names. ImprovMX blog (nofollow) for patterns only.
Confirm pricing. Free 1/3/14-day/HOLD/no send-as. Solo $40/yr. Starter $8/$80. Business $19/$190. Agency $39/$390. No SOC 2. No inbox SLA. No uptime SLA.
People-first launch notes name the zone.
The user problem and the decision criteria
Launch day is when leftover MX, theater NS, and self-send lies cluster. A checklist that says “test email” without naming the sender, resolver, and subject is decoration. The useful list is boring and dated.
| Question | If yes | If no |
|---|---|---|
| Two resolvers exclusive? | You may probe. | Stop the launch story. |
| Foreign sender? | Valid inbound proof. | Self-send. Invalid. |
| History 250 per noun? | Search the destination. | Classify empty vs fail. |
| Morning MX still exclusive? | You can close the window. | Registrar republished. |
Technical mail flow
Senders use public MX. Your probe must be a sender that uses public MX. Envelope SRS. Header From intact. IETF RFC 5321 — Simple Mail Transfer Protocol. If send-as is in scope, that is a second probe with dashboard SMTP and a mapped From—not a substitute for inbound.
TTL and cache exist. Two resolvers reduce “my laptop still sees Google.” They do not make dual MX safe.
Step-by-step setup and decision path
Sheet the zone
Nouns, destinations, NS, owner who can delete leftovers.
Build maps while old MX stays exclusive
Config parallel. Not dual MX.
Cut and screenshot exclusive MX
Filename zone + date. Two resolvers.
Foreign-probe every printed noun
Unique subjects. History. Destination search.
If send-as: one outbound 250
Paid plan. Mapped From. Third mailbox. Delete the script.
Morning re-query and HOLD review
Republish happens overnight. Write the date.
Failure modes and proof
| What you see | Likely cause | Proof |
|---|---|---|
| Green Slack, mixed MX | Self-send gate | Second resolver |
| One noun missing | Never created | Alias list |
| 250 empty inbox | Filter or HOLD | Recovery + search |
| Morning leftovers | Email tile republish | Re-delete + dated shot |
| 550 on send-as | Free or From | Plan, not the inbound checklist |
Proof is the dated pack: NS, MX, probes, morning re-query. “Felt fine” is not a pack.
MailerZ workflow and product boundary
We show history for hops we accepted. We do not promise inbox placement. We do not sell an uptime SLA. Envelope SRS. Header From untouched. Not IMAP. Not an open relay. Free 14-day store. Paid 90. Confirm /pricing. No SOC 2.
Cost, alternatives, and trade-offs
| Choice | What you get | What you give up |
|---|---|---|
| One-zone checklist | A closeable ticket | The Friday fleet fantasy |
| Paid monitor SaaS | Their probes. Quote live | Nothing if you skip leftover checks |
| Workspace cut playbook | A suite. Quote Google live | This hop if they stay on the suite |
Time is a line item. Thirty minutes of probes costs less than a Monday leftover.
Before the cut
Verify TXT already matches. Aliases created. Destinations 2FA. Old MX screenshot saved for rollback of the old world—and you accept that rollback is exclusive old or exclusive new, never mixed. Capacity counted. SMTP isolated if send-as is in scope.
During the cut
One zone. One operator pasting. One filename. Two resolvers before anyone sends a customer-facing “we’re live.” Unique probe subjects. Do not probe from the destination Gmail.
The next morning
Re-query MX. Re-open history for overnight 4xx. HOLD review. Website or registrar tiles overnight are the usual republish. IPv6 after exclusive names exist is a host ticket. Night operators who skip morning because “launch Slack was green” inherit leftover MX.
If a second operator needs the order, send this page plus leftover MX and recovery. Sheet. Exclusive. Foreign probe. Morning re-query.
The artifacts that close an email migration monitoring checklist are the dated pack. Everything else is a vibe you called a launch.
Worked scenarios for launch-day monitoring
A single-zone founder cut on a Tuesday. Two resolvers showed exclusive MailerZ MX. A friend sent a unique subject to hello@. History showed 250. Destination search found it. Next morning the MX set was unchanged. That is a finished checklist. Self-send from the founder's Gmail was never the gate.
An agency batched ten Webflow clients on a Friday. Nine probed clean. The tenth still had Google MX at priority 10 because a contractor enabled Workspace for Docs. Monday's donor hit Google. The missing sheet for zone ten is the whole postmortem. One zone, one dated packet. Batch mood is not a checklist.
A website save overnight republished the host's email product. Launch day looked exclusive. Saturday morning it did not. The checklist's morning re-query exists for this. Delete leftovers. Probe again. Assign the website owner so the next theme update does not undo DNS.
HOLD caught a customer who used support@ when only hello@ was mapped. History is not empty; the class is HOLD. Recover inside the store window if policy allows, or ask them to use the printed address. Do not enable catch-all FORWARD as a launch-day apology.
Send-as was in the launch deck on a Free plan. AUTH failed. That is not a monitoring miss. That is a plan miss. Confirm /pricing before the window. Do not add leftover MX because SMTP 550d.
A 'monitoring' Slack channel filled with full message bodies. That is a leak, not a checklist. Ticket IDs, subjects, hop codes, and resolver listings are enough. Bodies stay in the product or the destination.
Practice and anti-patterns on launch day
Practice: per-zone sheet with MX listings, probe IDs, leftover absent, morning owner. Anti-pattern: one Slack thread titled 'cuts' with ten screenshots and no filenames.
Practice: foreign sender. Anti-pattern: founder emails herself and screenshots Gmail.
Practice: unique subjects that include the zone and UTC date. Anti-pattern: 'test' sent twelve times.
Practice: two public resolvers, not only the DNS console. Anti-pattern: trusting the pencil icon.
Practice: leftover MX as a hard stop. Anti-pattern: 'we'll leave Google at 20 during QA.' QA is exclusive plus a probe.
Practice: morning re-query. Anti-pattern: closing the ticket at 11 p.m. because the UI looked right.
Practice: HOLD review if unknowns are on. Anti-pattern: ignoring HOLD and calling the launch failed.
Operator closeout before you leave the window
Closeout is a dated packet: two resolver listings, leftover MX absent, probe message IDs, Header From screenshot, destination search note, plan name if SMTP was in scope, morning owner.
Filename every screenshot with the zone and the date. 'img123' is how packets die in a Downloads folder.
If a probe failed, the packet names the class: leftover, HOLD, 5xx, 4xx, destination folder, or no send. Do not write 'weird' and go to dinner.
If SMTP was out of scope, the packet says so. Clients will still try Gmail send-as. They will fail on Free. That failure is documented, not a Sev-1.
Agencies do not invoice a batch closeout. One packet per zone or the tenth zone is fiction.
The drain clock for the old host is in the packet with the TTL you actually read. Then cancel the old seat. Courtesy without a clock is leftover MX next quarter.
Edge cases the checklist must name
Self-send looks delivered and the foreign probe is empty because leftover MX ate strangers. Believe the foreign probe. Fix exclusive MX.
Two brands, one launch call. The customer emailed the other zone. History on this zone is empty. Check the sister domain before you rebuild DNS.
DNSSEC SERVFAIL after a host move looks like a silent MX failure. Classify DNSSEC before you republish records in a panic.
A ticket tool ate the forwarded copy. MailerZ 250 is real. The Zap dropped it. That is not a launch-day MX miss. Check the filter log.
Null MX leftover from a brand pause travels into the new set if you exported sloppily. Review the listing. Delete null if you want mail.
IaC apply the next morning resurrects Google MX. The checklist must include repo grep, not only the console.
Field notes from launch windows
The calm launches are boring packets. The loud launches are heroes on Slack. Hire for boring.
Agencies that bill a monitoring hour have fewer Monday donor tickets. Agencies that skip it donate the weekend.
UTC on the sheet beats 'this afternoon.' Destination search on the wrong day is a self-own.
Use /tools and /troubleshooting when the console and the world disagree. The console is one view. RFC 1035 is what senders implement.
No inboxing percentage on the launch email to the client. You can promise a hop and a leftover check. You cannot promise a tab.
Website teams will republish email products. The morning row exists because theme saves do not read your runbook.
Handoff memo for the morning owner
Zone, registrar, DNS panel, MailerZ workspace, printed aliases, HOLD policy, where history lives, and who may recover HOLD. If the morning person cannot open those, they will flap MX.
The memo forbids self-send as proof. It forbids adding the old host 'until we're sure.' Sure is exclusive plus a foreign probe.
SMTP scope is written. Dashboard copy only. No folklore ports. Free has no send-as.
Link /email-forwarding and /docs. Inbound proof is not send-as.
If a website owner can publish DNS, they are in the memo. Otherwise the morning re-query will surprise you every launch.
Acceptance criteria for a launch-day checklist
Two resolvers exclusive. Leftovers gone. Foreign probe per printed alias. History class written. Destination search done. Morning re-query scheduled with a name.
Screenshots filed with zone and date. No anonymous Downloads pile.
Self-send was not the gate. Batch cuts have N packets, not one vibe.
HOLD and SMTP plan match what you sold. No panic FORWARD. No invented SLA.
IaC and console agree, or you documented console-only and locked the module.
No customer bodies in Slack. Ticket IDs only.
Operations review of launch monitoring
Count last quarter's launches that skipped morning re-query. If leftovers returned, that skip is the cause. Make the morning row blocking for invoice.
Count leftover-MX tickets in week one. If the number is not zero, the pre-cut listing step is theater.
Review who holds registrar 2FA. A launch without that person is a launch you cannot roll back.
Review filename discipline. Packets you cannot find are packets you did not keep.
Quote /pricing before the next window if SMTP is newly in scope. Do not discover Free mid-cut.
Rehearse on a lab zone. The checkout domain is the wrong first checklist.
Quarterly review of the checklist itself
Did provider UIs change? Update screenshots in the internal wiki, not only this article.
Did MailerZ MX hostnames change? Copy the dashboard again. Memory is leftover's cousin.
Did we batch without sheets? Write the miss. Do not 'try to be faster' next time.
Did we promise inboxing in a launch email? Retract the macro.
Did website wizards return? Assign the owner. Disable the product that rewrites MX.
Author: MailerZ editorial, Secuno LLC. Review when DNS hosts, pricing, or scope change.
Closing notes on watching the cut
A launch-day email monitoring checklist is two resolvers, a foreign probe, hop history, leftover MX absent, and a named morning re-query—per zone. It is not a self-send and not a Friday batch mood.
Filename the proof. Classify failures. Keep bodies out of Slack. Exclusive MX is the launch. The checklist is how you know it stayed launched.
Start free on a lab, run the sheet once, then cut production. Product path: /tools, /troubleshooting, /email-forwarding.
People and deputies on launch day
Name the cutter, the reviewer who reads listings, the morning owner, and the website owner who can republish MX. If the website owner is missing, the morning row will keep failing. Launch-day monitoring is a staffing diagram, not only a DNS trick.
A junior can run the sheet if the sheet is written. A junior cannot invent two-resolver proof from a Slack vibe. Pair them once on a lab zone. The first production cut should not be their first packet.
Agencies need a client approver who will not demand dual MX 'for QA.' If that person is not in the call, the call will produce leftover MX. Invite them or refuse the dual-publish request in writing before the window.
Support must not promise inboxing while you watch hops. Give them the class list: leftover, HOLD, 5xx, 4xx, folder, no send. They classify. They do not apologize with a DNS flap.
If nobody owns filenames, the packet lands in Downloads as img_final_final. Assign filenames. Zone plus UTC date. That is a person assignment.
Proof packet contents that close the window
Two public resolver listings, leftover MX absent, unique foreign-probe subjects, history codes, destination search notes, HOLD review if unknowns exist, SMTP plan if send-as was in scope, morning owner name, IaC grep if you use it.
Screenshots are evidence when they include the zone name in the filename and the resolver name in the image. A crop of a green check without context is decoration.
Self-send screenshots are allowed as a curiosity. They are not the gate. If the packet's only pass is a self-send, the packet fails this checklist.
If a probe failed, the packet holds one class, not five simultaneous edits. The next edit waits on the class. That sentence keeps launch day from becoming a rewrite of the whole internet.
Bodies stay out. Subjects, IDs, codes. If counsel wants a body, recover inside the store window under policy. Slack is not the store.
Appendix: launch myths that waste the window
Myth: TTL is always twenty-four hours. Read the number. Wait that number plus a buffer before you cancel an old seat.
Myth: priority 20 on Google is safe QA. If it answers, it is leftover MX. Delete.
Myth: the DNS console is the internet. Query two public resolvers. RFC 1035 is what senders use.
Myth: ten zones on one Friday is efficient. Ten packets is efficient. One vibe is a Monday donor.
Myth: morning re-query is optional if last night looked good. Website wizards work while you sleep.
Myth: send-as on Free will work if the checklist is good. Free has no send-as. Confirm /pricing before the window, not during AUTH failure.
More operator notes for the hour after cut
Do not stack a registrar NS change, a website redesign, and an MX cut in one hour. Each class has its own failure mode. Unstack. The checklist assumes email is the change you are watching.
If history is empty and leftovers are gone, ask whether they sent. Sent-folder proof is kindness. Empty history plus a draft is not a MailerZ incident.
If history is 250 and the owner cannot find the message, search all mail. Then spam. Then the exact subject. Folder class. Not a republish.
If HOLD has the copy, decide recover versus resend. Do not enable catch-all FORWARD as fireworks to celebrate the launch.
Write the next quarter's review date on the packet. Plans and leftover products change. A launch without a review date is a eulogy.
Product path: /tools, /troubleshooting, /email-forwarding, /docs. Author: MailerZ editorial, Secuno LLC. No inboxing percent. No badges.
Extended narrative of a single-zone launch
Imagine one zone, one printed hello@, one destination Gmail, and a Friday window you actually control. Monday you write the sheet. Tuesday you confirm the map exists and old MX is still exclusive so you have a rollback picture. Wednesday you lower TTL if the record lets you. Thursday you copy MailerZ MX from the dashboard, not from last quarter's screenshot. Friday you publish exclusive, delete leftovers, query two resolvers, and ask a friend to send a subject that includes the zone and the UTC date.
History shows 250. The founder searches all mail and finds it. You do not let them email themselves as the gate. You filename the listings. You name Saturday's owner—often you. You go home. Saturday you re-query. If a website wizard republished Google MX, you delete it again and probe again. That Saturday row is the difference between a launch and a coin flip that waited until a customer found it.
If HOLD has a note from someone who used help@, you either recover under policy or you ask them to use hello@. You do not enable catch-all FORWARD as champagne. If SMTP was on the slide deck and the plan is Free, you do not discover that during AUTH. You discovered it Thursday when you read /pricing. The 550 is not a monitoring failure. It is a scope failure you refused to stack into MX.
Now imagine ten zones with no sheets. You will remember nine. The tenth is the donor. This narrative exists so you refuse the batch mood. One zone. One packet. One morning name. Then the next zone, if the first packet is closed.
Agencies bill this hour because the unpaid version still happens, only later, with worse screenshots. Founders do it once and keep the packet next to the registrar 2FA. That is a launch-day email monitoring checklist when it is real, not when it is a slogan on a standup.
Endnote on watching instead of hoping
Hope is a self-send. Watching is two resolvers, a foreign probe, a leftover check, and a morning name. MailerZ gives hops and a short store. It does not give you an inboxing percent or a badge. Use the hops. File the packet. Do not dual-publish to feel safe.
Start free on a lab and run the sheet once before production. Product path: /tools, /troubleshooting, /email-forwarding. Author: MailerZ editorial, Secuno LLC. Review the checklist when DNS hosts, website builders, or pricing change.
Last checks before you call the cut done
Read the sheet out loud. Two resolvers. Leftover absent. Foreign probe IDs. Destination search. Morning name. If any line is a shrug, the cut is not done. A shrug is how leftover Google MX survives a Friday.
Confirm SMTP scope against /pricing. Free has no send-as. A 550 during the window is a scope miss, not a reason to republish the old host. Do not stack AUTH into MX.
Confirm IaC will not resurrect old MX on Monday. Grep the repo. The console and the module must agree. Otherwise the morning row is theater.
Bodies stayed out of Slack. Filenames have the zone. The packet is findable. That is the last check. Then you may cut the next zone, not before.
FAQ
What is the safest way to handle email migration monitoring checklist?
Before cut: maps exist, old MX still exclusive. At cut: MailerZ MX only, leftovers deleted, two resolvers agree. After cut: foreign probe every printed string, history 250, destination search, leftover check. Next morning: re-query MX. Do not use self-send as the gate. Do not batch-cut ten zones without a per-zone sheet.
Does this require a new mailbox?
No. The probe destination can be the existing Gmail. Use a third mailbox as the sender. MailerZ is not IMAP.
Will it work with Gmail or Outlook?
Yes when MX is exclusive and aliases exist. Gmail self-send can lie. Outlook as destination still needs a foreign sender. Free has no send-as. Confirm /pricing.
What DNS records are involved?
Exclusive MailerZ MX, verification TXT already green, leftovers gone. SPF/DKIM/DMARC if send-as is in the launch. See RFC 5321 and RFC 1035.
What should I test before production?
This page is the test. Print the checklist. Run it on one zone. Keep the dated screenshots. Then you may cut the next zone.
Key takeaways
- Two resolvers. Exclusive MX. Zone-named screenshots.
- Foreign probe every printed noun.
- History 250 plus destination search.
- Self-send is not the gate.
- Morning re-query for republish.
- One zone per window.
- Send-as is a second probe. Free has none.
- Confirm /pricing. Not an inbox SLA. Not SOC 2.
Conclusion and next action
If you cut MX, monitor listings and hops, not feelings. MailerZ can show history for mail it accepted. It cannot watch a leftover you left in the panel. Start free on one domain and run the checklist once for real.
Ready to monitor one cut
Start free with one domain and a dated launch pack.
Two resolvers. Foreign probe. Morning re-query. Sign in if the domain is already there.
Review after each cut style changes and quarterly. Author: MailerZ editorial, Secuno LLC.