IP address warm-up

What is IP warm-up and why it is needed

Warm-up is the gradual increase of the sending volume from a new IP address to build a positive reputation with mail providers (Gmail, Yahoo, Outlook). Providers use the IP address to identify the sender, track its behavior, and assign a reputation.

An IP address that is new to your infrastructure may have no sending history; it may also have been used by another sender before. If you start mass sending right away, providers may throttle acceptance, reject messages, or route them to spam. In Altcraft Platform, warm-up is performed by the AKMTA sender: the volume is controlled by sending strategies, and the provider response — by retry and lock rules.

Preparation

Before starting the warm-up, configure the sending infrastructure:

• DNS records — for each sender IP address, create an A record on the sender domain, set up PTR, SPF, DKIM, and DMARC records. The step-by-step scheme is described in Email: ISP interaction guidelines;
• IP addresses on the node — the addresses must be available on the sending node and added to the AKMTA sender settings;
• DKIM and pools — DKIM keys are created in the admin panel; when sending from different domains, use pools;
• Sending service — if you use external services (for example, Amazon SES instead of or alongside AKMTA), verify the domain and lift the sandbox restrictions: in Amazon SES, you need to obtain production access separately;
• Clean list — send only to confirmed subscriptions; addresses with hard bounces are added by the platform to the global stop list;
• List segmentation — split the list from the most active subscribers to the least active and start with the active segments.

Gmail's TLS, DMARC alignment, and one-click unsubscribe requirements for bulk mailings are listed in the Gmail sender guidelines.

Volume control via AKMTA strategies

Build the warm-up plan based on your target volume, mailing regularity, and the distribution of recipients across mailbox providers. Increase the volume after evaluating the results of previous sendings; pause the increase if bounces, deferrals, or complaints grow.

In the platform, the daily volume is set via sending strategies: the product of Concurrent connections, Messages per connection, and Time interval determines the speed, and the Average strategy speed per IP address table shows the estimated volume per minute, hour, and day for a single IP. For warm-up, create a separate strategy with a low speed and raise it in steps as the metrics stabilize.

The speed is configured separately for each ISP: the Speed rules in the sender settings let you set different speeds for IP address and provider combinations, including Gmail, Yandex, Mail.ru, and Outlook.com. This is useful when providers respond to warm-up differently.

note

Ready-made warm-up schedules depend on the sending service. For example, the generic Twilio SendGrid guide contains a 21-day plan for a target volume of 1,000,000 messages per day, while automatic warmup of standard dedicated IP addresses in Amazon SES is designed for 45 days. Use them as a reference for your own strategy, not as a universal rule.

Retry and lock rules

The response to provider rejections during warm-up is controlled by AKMTA rules:

• Retry rules — if the provider does not accept a message immediately (a temporary 4xx rejection), the platform retries the sending. Each rule defines the retry mode (same IP, another IP of the same node, any IP of any node), the number of retries — no more than 10, the interval, and its growth type: fixed, geometric, or random;
• Lock rules — after sustained negative responses, the platform suspends sending to the provider entirely or from the current IP address for a specified time.

When the retries are exhausted or the message retention period in the sender queue expires (three days by default; the value is set in the platform configuration), the message is registered as undelivered — with the reason shown in the undeliveries report.

For warm-up this means: temporary deferrals do not stop the sending — messages are retried according to the retry rules, while a provider's sustained rejections are taken out of sending by the lock rules.

Monitoring

During the warm-up, track the indicators in the platform reports:

Metric	Where to look in Altcraft
Deliverability and undeliveries	Undeliveries report — reasons per campaign
Bounces	Bounces report: the platform records hard bounces in the stop list and retries soft bounces by the retry rules. The guideline for hard bounces is up to 0.5% of sent messages (ISP interaction guidelines)
Complaints	The Complaints and Complaint rate metrics in the channels report. Internal guideline: on average no more than 0.07%, with 0.1% as the limit; for Gmail, watch the spam rate in Postmaster Tools — 0.3% and higher is the risk zone
Postmasters	External provider services — postmaster setup

caution

If the bounce rate or complaint rate grows — slow down or stop the warm-up and find the cause. To do this, lower the speed in the sending strategy and resume sending at a reduced volume, increasing it after the metrics stabilize.

To monitor reputation, use the provider tools: Google Postmaster Tools, Mail.ru Postmaster, Yahoo Sender Hub and CFL, Microsoft SNDS — the setup order is described in the postmaster setup article. FBL complaint reports are received by the incoming mail server built into AKMTA: it recognizes ARF reports in the abuse@, abusemaster@, and fbl@ mailboxes and registers the complaint in campaign statistics. The setup scheme is described in Email: ISP interaction guidelines.

Multiple IP addresses

• Each new IP address is warmed up independently: the speed strategy is set per IP;
• If you already have warmed-up IPs — they keep working while the new one is warming up: the distribution across IPs is determined by the sender Speed rules;
• Do not add IP addresses without need and do not use switching between them to bypass blocks: providers recognize the "snowshoeing" tactic — spreading spam across many IP addresses and domains (Spamhaus glossary). Each IP address needs its own warm-up plan and regular traffic.

Transactional and marketing campaigns

• Marketing campaigns: strictly follow the warm-up plan and raise the speed in steps;
• Transactional messages: the volume is not controlled by the schedule — it depends on user actions. When moving to a new IP address, migrate the flow gradually, and for urgent messages arrange sending through already running infrastructure in advance. Keep monitoring bounce and complaint metrics.

What not to do

• Do not send to the entire list at once;
• Do not use purchased or outdated email lists;
• Do not change the IP address in the middle of the warm-up;
• Do not ignore complaints and bounces;
• Do not expand sending to less active segments until the metrics stabilize.