Regulatory shifts, post-mortems, deep technical work, and the operational realities of running anonymous hosting for email senders. Long-form, narrative, written by operators who run the infrastructure.
DKIM 1024-bit keys remain widely deployed despite the migration recommendation toward 2048-bit having been in place for years. We finished migrating our entire customer base to 2048-bit DKIM signing in early 2026. What the migration involved, what we learned, and what other operators should know.
Some operations have privacy requirements that go beyond standard business practices. Journalists, legal advocacy, healthcare-adjacent operations, opposition political work. The email infrastructure for these operations differs from standard practices. What matters and what does not.
Yahoo started filtering mail from multiple customer accounts to spam folder simultaneously on a Tuesday in early April. The pattern was distinct enough to suggest receiver-side change rather than coincident sender-side issues. The investigation, the suspected cause, and how we mitigated.
The 'spam filter' that senders worry about is not a single mechanism. Modern receivers run dozens of evaluation systems in parallel. Understanding the major categories of evaluation helps senders think clearly about what they can and cannot influence.
Microsoft's email authentication evolution continues. The latest update affects how Outlook handles ARC-sealed mail and what intermediaries must do to maintain delivery quality. The specific requirements, the trusted sealer concept, and what senders need to know.
Our BTCPay Server installation stopped sending payment confirmations to customers for 14 hours on a Friday. The cause was non-obvious. The diagnostic involved Bitcoin node sync issues, NBXplorer state, and webhook delivery. The full post-mortem and what we changed.
DMARC aggregate reports arrive as gzipped XML files daily. Most operators forward them to a third-party service for parsing. The reports contain operational information that third-party dashboards summarize but obscure. The XML schema, what the fields mean, and what you can find when you parse them yourself.
Customer payments arrive in Bitcoin; we prefer accounting in Monero for privacy-aligned operations. Atomic swaps replace the centralized exchange dependency. The current state of BTC↔XMR atomic swap tooling, what works, what does not, and our production experience.
Three major enforcement events in 2025. Microsoft enforcement in May, Postmaster Tools v1 retired in September, Gmail hardening in November. The year's events and what they mean cumulatively for senders heading into 2026.
Forwarding scenarios break DMARC alignment. Mailing lists add headers and footers that invalidate DKIM signatures. ARC was designed to solve this. Implementing it correctly takes more than installing OpenARC. The architecture, the configuration, and what we have learned about ARC in production.
Cold outreach via direct SMTP became operationally untenable through 2024-2025. The multi-mailbox approach (sending through Google Workspace and Microsoft 365 inboxes) emerged as the working alternative. The architecture, the operational reality, and what is and is not viable in late 2025.
Google updated their Bulk Sender Guidelines FAQ on November 3, 2025 announcing that enforcement is ramping up. Non-compliant mail now experiences temporary and permanent rejections at SMTP level. The error codes, the patterns from the first week, and what this means.
A customer's SendGrid account got terminated on a Tuesday with three days until their largest annual campaign. We migrated them to self-hosted infrastructure in 72 hours. The decisions made under time pressure, the corners we cut, the corners we did not, and what we learned about emergency migration.
Google retires Postmaster Tools v1 on September 30, 2025. The Domain and IP Reputation dashboards go away permanently. Our customer migration through v2 over the past year revealed specific patterns, surprises, and what to monitor without the reputation graphs.
We were calling CoinGecko's API every five minutes to display BTC and XMR prices on our pricing page. Rate limits, dependency exposure, and operational drag motivated us to build our own self-hosted price aggregator. The architecture, the trade-offs, and the open-source code we released.
Three self-hosted email marketing platforms serve overlapping but distinct operator profiles. MailWizz dominates the commercial self-hosted market. Acelle offers an alternative with different architecture. Sendy focuses on Amazon SES integration. Comparison from operator perspective.
TLS-RPT reports are a niche feature that most operators ignore. We have been parsing them across customer infrastructure for two years. The patterns in 100,000 reports reveal infrastructure problems senders rarely diagnose. The findings and what to do with the data.
A major mailbox provider had a 6-hour outage on a Thursday morning. Our customer queues grew accordingly. The expected behavior, the surprising operational decisions we faced, and what we learned about PowerMTA queue management during multi-hour external incidents.
We enabled a strict Content-Security-Policy on our marketing site for visitor privacy. Cloudflare's Email Obfuscation feature injects inline JavaScript that violates strict CSP. The diagnostic took us four hours to identify. The fix took three minutes. The procedural lesson took longer.
Transactional and marketing email share infrastructure but require different operational disciplines. Customers conflating them produce surprising outcomes when transactional volume scales. The differences that matter and the architectural decisions that follow.
Microsoft's bulk sender enforcement started May 5th with hard rejection rather than the originally announced quarantine. The first week produced specific patterns across our customer base. Specific numbers, what is breaking, and what is working differently than expected.
Microsoft published their bulk sender requirements on April 2nd with enforcement starting May 5th. Three weeks of preparation time for an enforcement comparable to Gmail's February 2024 deadline. What is required, what is different from Gmail, and what to do this week.
PTR records and reverse DNS look like infrastructure plumbing nobody needs to think about. Then mail starts landing in spam folders for inscrutable reasons and the investigation eventually surfaces FCrDNS configuration as the cause. The mechanics, the failure modes, and the operational fix.
The Promotions folder in Gmail is not a punishment. It is a classification. Understanding which classification Gmail applies to your mail and how to influence it requires understanding signals senders rarely look at. The mechanics of inbox vs Promotions sorting and what we have learned about influencing it.
The IP warming schedules circulated in older deliverability documentation no longer match how receivers evaluate sender behavior. The 2025 receiver model uses engagement-weighted signals that change what warmup needs to look like. The schedules we use now and why.
Several customers asked us this year about building their own ESP business on infrastructure we provide. The economics work for specific operator profiles. The six-month operational curve is what catches new ESP builders by surprise. The realistic timeline, capital requirements, and what most teams underestimate.
An IP that looks clean in five seconds of checks can be radioactive when you look harder. This is the 14-check pre-flight we run on every IP before assigning it to a customer, what each check catches, and the rate at which IPs fail each one.
A scheduled DKIM key rotation produced 48 hours of customer email delivery failure. The cause was procedural rather than technical. The €18,000 estimated business impact and what we changed about our rotation procedures afterward.
BIMI promised brand logo display in Gmail and Yahoo. The VMC certificate requirement changes the calculation. The €1,500/year cost, the trademark verification process, and whether BIMI is worth pursuing for senders below enterprise scale.
Migration from Mailchimp to self-hosted infrastructure is operationally bounded but rarely as fast as customers expect. The 200K subscriber threshold is where the economics make sense and the engineering work becomes worthwhile. The real timeline, cost breakdown, and what we have seen go wrong.
PowerMTA's queue system is straightforward until something goes wrong. Then the queue states, retry logic, and backoff behaviors become operationally important. This is the field guide based on production troubleshooting of stuck queues across multiple customer deployments.
June 1st was the soft deadline for the List-Unsubscribe-Post header implementation per RFC 8058. We migrated our customers ahead of the deadline. Two specific implementation problems caused production issues. The fixes and what to verify in your own setup.
A customer's IP range appeared on Spamhaus DROP list for approximately 8 hours due to compromise of a different operator sharing the upstream provider. The remediation involved provider coordination, evidence gathering, and verification across multiple systems. The full timeline and what we changed.
A customer's IP hit the Spamhaus CSS list on a Tuesday. The remediation took nine days from initial detection to confirmed delisting. The sequence reveals what works and what wastes time when working with Spamhaus listings.
Gmail's enforcement tightened in late April. The 5% rejection threshold for non-compliant senders is producing measurable impact. Three operators across different segments shared their week-over-week numbers. The patterns tell the story of where the rules are biting hardest.
Google released the Postmaster Tools v2 dashboard on March 18th. The new compliance section maps directly to the February enforcement model and tells senders exactly where they stand. What the new dashboard actually shows and how to read it.
MTA-STS (RFC 8461) provides policy-based enforcement of TLS for SMTP. Implementation is bounded but has specific pitfalls. The deployment process, the operational considerations, and what the policy actually accomplishes for sender and receiver.
The three primary MTA choices for self-hosted email operations serve different operator profiles. PowerMTA dominates commercial high-volume sending. Postfix handles general-purpose mail. Postal offers open-source alternative to PowerMTA. Operator perspective on each.
Gmail started enforcing the February 2024 bulk sender requirements one week ago. The numbers from our customer base show a pattern very different from the noisy predictions. Specific deliverability data, error patterns, and what is working versus what is breaking.
The February 2024 Gmail and Yahoo deadline arrives in less than two weeks. This is the audit sequence we ran across our customer base and what the most common failures looked like in the field.
We added Lightning Network as a payment option in early 2023. After a full year of production use, the operational characteristics, customer adoption patterns, and what we learned about running a Lightning node for hosting invoice payments.
Four of the major SaaS ESPs serve different operational segments well. We have worked with customers across all four. The honest comparison from an operator perspective, what each does well, where each falls short, and which fits which use case.
The DMARC progression from monitor to enforcement is operationally fraught. Skipped phases produce lost mail. Rushed timelines produce reputation damage. This is the eight-month sequence we use with customers and what we have learned about the failure modes.
Gmail and Yahoo announced new bulk sender requirements last week. Most of the analysis online is missing what matters operationally. Here is what we are doing about it before February.
Customers crossing certain volume thresholds find the SaaS ESP economics no longer competitive with self-hosted alternatives. The migration path involves technical work, organizational change, and operational learning. What works, what does not, and how to think about the transition.
The SPF 10-lookup limit silently breaks email for senders growing through SaaS adoption. This is the operational guide to detecting, debugging, and fixing PermError scenarios that producers spend years not realizing they have.
On a Thursday in June, our upstream provider in Bucharest had a BGP session bounce twice in twelve hours. The flap itself lasted seconds. The deliverability damage took three weeks to fully unwind. A walk through what happened, what we learned, and what we changed.
DKIM rotation is one of those tasks that looks simple in documentation and breaks in subtle ways in production. A complete walk through the operational sequence, what we have seen go wrong, and the timing nobody warns you about.
We have been running BTCPay Server in production for accepting cryptocurrency payments since 2022. The setup, the operational reality, and what the alternative would have cost. This is what works for our specific use case.
Most cold outreach agencies are paying for infrastructure they do not need and missing the pieces they cannot operate without. A walk through what the real operational requirements look like at the agency volume tier.