Skip to content

Monitoring Your Email Infrastructure

Monitoring 9 min read
On this page

"Our mail stopped being delivered" is rarely one failure. The mail host can be down. A sending IP can be sitting on a spam blacklist, an MX or SPF record can have been changed under you, or a record can have been wrong since the day it was published. Uptimia's monitors watch the first three continuously. For the fourth, a Security monitor reads your SPF, DMARC and mail encryption setup once a day and reports the common mistakes. Anything beyond those is yours to check before you publish the record.

Before You Start #

  • The Free plan allows 0 blacklist, 0 DNS and 0 Security monitors. Basic includes 1 of each, Professional 10, Enterprise 100. Uptime monitors, including the Email sub-type, are on every plan. The full per-plan counts are in Plan Limits and Quotas.
  • You need the Editor role or above to create monitors. Read-only members can see them and receive their alerts.

What Breaks Email Delivery #

What breaksWhat you seeThe check that catches it
The mail host stops answeringQueued mail, connection timeouts, no bounceUptime monitor, Email sub-type, on SMTP / POP3 / IMAP
A sending IP or your domain lands on a spam blacklistMail accepted by you, rejected or filed as spam by themBlacklist monitor on the domain and every sending IP
An MX, SPF, DMARC or CAA record changesMail routed elsewhere, or suddenly failing authenticationDNS monitor on the zone, mail and policy groups
A record was wrong from the startNothing fires: the record never workedSecurity monitor, for missing SPF or DMARC, +all, p=none, no STARTTLS and no MTA-STS. Check anything else yourself before you launch

The first three run continuously. A record that was never right does not change, so the DNS monitor has nothing to report about it. The Security monitor judges the record itself, but only for the mistakes listed under Is the Record Right? below. Read the record yourself before you publish it, and again after every edit. That is what catches an SPF record that needs more DNS lookups than the limit of 10 allows, which no monitor checks.

Is the Mail Server Answering? #

Create one Uptime monitor per mail host. On the Main tab, set What should we check? to Email, then pick the Protocol: SMTP, POP3 or IMAP. Enter the host in IP / Domain and the port in Port. Turn on Use StartTLS (SMTP) or Use SSL/TLS (POP3 and IMAP) if the port expects encryption.

Point the host at the MX hostname itself. Your web domain frequently resolves to a different machine, and a monitor on the web domain will stay green through a mail outage.

Two fields turn the check from "the port is open" into "the mail server itself answered". String to expect works on all three protocols. String to send is SMTP only. An SMTP greeting begins with 220, so 220 in String to expect proves a real server answered rather than a load balancer accepting the connection. The check log records both fields as String sent and String expected.

The uptime editor with the Email sub-type selected, showing the SMTP protocol choice, host and port fields, the string-to-send and string-to-expect fields and the StartTLS toggle.Enlarge

Everything else about this monitor is the ordinary uptime family: check frequency, alert delay, probe locations, alert recipients. All of it is documented in Uptime Monitoring. The keyword check is a Web sub-type control that does not render on Email, so String to expect is its equivalent here.

Are You Listed? #

A Blacklist monitor sweeps 17 DNSBL zones for one domain and the addresses that send for it. Enter a naked domain in Domain or IPv4 and it discovers the rest on every sweep. The web IP comes from the A record, and the mail-server IP from your main MX record (the one with the lowest preference number) resolved to its A record.

The sweep misses any sending IP that lives outside your DNS, such as the dedicated address your email service provider (ESP) gave you. Add those under Dedicated sending IPs in the What we watch card, up to 5.

The What we watch card on a blacklist monitor, showing the domain-to-website-IP-and-mail-IP fact line and two dedicated sending IPs added below it.Enlarge

One monitor covers one domain. If you send from several domains, each needs its own. Accepted targets, sweep cadence and tier configuration are all in Blacklist Monitoring.

Reading a Listing: Major, Standard and Informational #

The 17 lists sit in three tiers, and the tier decides how loudly a listing alerts.

TierListsDefault severity
MajorSpamhaus ZEN, Spamhaus DBL, BarracudaCritical
Standard12 lists, including SpamCop, SURBL, URIBL, PSBL and SpamRatsDegraded
InformationalUCEPROTECT L1 and L3None (counted on the monitor, never opens an incident)

Large providers and appliances consult the three Major lists directly, so a listing there bounces real mail. The Standard lists are scoring signals for SpamAssassin-class filters. One listing nudges a spam score; several compound.

Each tier has its own toggle and, for Major and Standard, a Critical / Degraded severity select in the Lists & alerts card.

The Lists and alerts card on a blacklist monitor, showing the Major and Standard tiers with their Critical and Degraded severity selects, the Informational tier marked recorded and never paging, the sweep interval and the clean-answer streak that closes a listing.Enlarge

One clean answer does not clear a listing. Treat a listing as resolved after offers 1, 3 or 5 consecutive clean answers, and defaults to 3, which is 30 minutes of agreement at the default 15-minute sweep. Blacklist Listings and Delisting covers the lifecycle, the per-list delisting routes, and what happens when a listing closes without ever being verified.

Did Your Records Change? #

A DNS monitor watches one zone at its authoritative nameservers and compares every answer against a baseline. Two of its four record groups carry mail:

GroupRecordsDefault severity
MX@ MX (mail routing)Critical
Policy records@ TXT (SPF), _dmarc TXT, @ CAAWarning

The other two groups, Apex & host addresses and NS, matter to mail indirectly. The blacklist monitor finds your web IP in the A record, and an NS change moves the whole zone.

An MX change reroutes live mail, so it pages at Critical. An SPF or DMARC edit degrades authentication over hours, so it opens a Warning incident instead. Each group's select offers Critical, Warning and Info only, and you can change either severity on the group itself. Setup, nameserver health and the change workflow are in DNS Monitoring.

The Watched records card on a DNS monitor, showing the four record groups with MX set to Critical and Policy records set to Warning.Enlarge

Is the Record Right? #

A Security monitor on your website's URL also runs three mail checks on its domain, once a day. In the monitor's Checks card they are Email authentication (SPF and DMARC), Mail encryption (STARTTLS on your MX hosts) and MTA-STS. All three are on by default.

The Checks card in the Security monitor editor with the Important and Hardening tiers opened, showing the Email authentication, Mail encryption and MTA-STS switches among the other check groups, and the Alert on select below them.Enlarge
FindingSeverity
The SPF record ends in +allHigh
A mail server does not offer STARTTLS, or the handshake failsHigh (shown on the monitor page only, no alert for now)
No SPF recordMedium
No DMARC record, an invalid one, or more than oneMedium
The DMARC policy is p=none, or the record has no policyLow
A domain with MX has no MTA-STS recordLow

Every finding is listed in the monitor's Open issues card. Only High and Critical findings can alert, and only when Alert on is set to High and critical. The default, Critical only, alerts on none of the mail findings. A Security incident reads Warning and never counts as downtime. Setup, the other check groups and Mark as expected are in Security Monitoring.

IPv4 Only: What a DNSBL Cannot Tell You #

DNSBLs are IPv4-only. A blacklist monitor accepts a naked domain or a bare public IPv4 address. It rejects everything else with the message "Enter a naked domain or a public IPv4 address": IPv6 addresses, RFC 1918 private ranges such as 10.0.0.5 and 192.168.1.1, loopback, CGNAT (addresses an ISP shares between customers) and the documentation ranges. When a watched address exists only as an IPv6 (AAAA) record, the monitor keeps the column in the results table. Every IP-list cell then reads not checked (IPv6) rather than an all-clear the monitor never measured.

If your mail host is IPv6-only, the blacklist monitor cannot answer for it, though the Uptime check and the DNS record watch still can. When a slot resolves to several IPv4 addresses, the monitor watches only the lowest one. Add the others as dedicated sending IPs if they send mail.

Who Gets Paged, and Where #

A blacklist incident and a mail-server outage need different people, and different urgency. You set recipients per monitor, so keep the two apart.

  • The Uptime monitor on the mail host is a page. Mail is queuing while it is down. Send it wherever your outage alerts go, and give it the alert delay you give any other production host.
  • The Blacklist monitor alerts the moment the incident opens. There is no alert delay control: a listing is already a fact by the time a sweep sees it. Delisting is deliberate work, and it belongs to whoever owns sending reputation rather than whoever is on call at 03:00. Route it to that person or to a mail-ops channel.
  • The Informational tier never opens an incident, so it never pages anyone. The monitor still counts it, so you can see it.
  • The Security monitor alerts only on findings at or above its Alert on level, and its incidents never count as downtime. Send it to whoever owns the DNS zone.
  • A blacklist monitor cannot be added to a status page. There is no status-page selection on the family, so a listing is never published to your visitors.

The recovery alert, Alert when every listing is verified clean, is on by default and fires only after the clean-answer streak completes.

One domain, mail on its own host, one dedicated sending IP from an email service provider:

  1. Verify the MX, SPF, DKIM and DMARC records themselves first, and fix anything wrong. The monitors then start from a healthy baseline.
  2. Create an Uptime monitor: Email, SMTP, host mail.example.com, port 587, Use StartTLS on, 220 in String to expect.
  3. Create a Blacklist monitor on example.com and add the ESP's sending IP under Dedicated sending IPs.
  4. Create a DNS monitor on example.com, leaving MX at Critical and Policy records at Warning.
  5. Create a Security monitor on https://example.com and open its Open issues card after the first scan for the mail findings.
  6. Point the uptime monitor's recipients at your on-call channel and the blacklist monitor's at whoever owns sending reputation.

Note: On Basic, steps 3 and 4 consume the entire blacklist and DNS allowance, 1 of each. A second sending domain needs Professional. The Security monitor in step 5 uses its own allowance, also 1 on Basic.

Was this helpful?

Your answer helps the team improve this article.

Still need help?

Uptimia Bot answers right away and passes you to the team when it can't.

Source: https://help.uptimia.com/help/a/monitoring-your-email-infrastructure