VMTech
Discuss a project →
← All capabilities
OPERATIONS & COMMUNICATIONS · 14

Server maintenance

A server gets remembered on two occasions: when it goes down and when the mail lands in spam. Our job is that neither happens — and if something does, it is fixed before a client notices.

Discuss a project
AN INCIDENT

What happens at night when a disk starts filling up

The real sequence: from the first warning to the report entry — with no call from the client.

monitor.vmtech.rs
Node: web-01
LIVE
18 minutes
A threshold is crossed disk 85%
The on-call engineer is paged 02:41
The cause is found application logs
Resolved disk 34%
Services verified all services healthy
Logged in the report 18 minutes
AN INCIDENT
  1. A threshold is crossedMonitoring sees the disk filling and warns early, not once everything has stopped
  2. The on-call engineer is pagedA message reaches the engineer with the node, the metric and the recent changes
  3. The cause is foundLogs show what is growing: the mail queue, application logs or temporary files
  4. ResolvedSpace is freed and log rotation fixed so it does not return in a month
  5. Services verifiedThe site, the mail and the data exchange are checked — not just “the server responds”
  6. Logged in the reportThe monthly report shows what happened, how long it took and what was done to prevent a repeat
A FAMILIAR PICTURE

What “the server is fine” usually means

Without maintenance

  • You learn about an outage from a customer or a message saying “your site is down”
  • Backups are supposedly running, but nobody has ever tried a restore
  • Mail from the domain lands in spam and nobody knows why
  • Updates are skipped because “if it works, don't touch it”
  • A former contractor still has server access, and nobody remembers who exactly

Under maintenance

  • Monitoring warns before failure: disk, memory, load, certificate expiry
  • Backups are verified by restoring them — the recovery time is a known number
  • The mail domain is configured: SPF, DKIM and DMARC in place, reputation watched
  • Security updates are applied on a schedule, in a window, with a rollback path
  • The access list is known and reviewed — a departed contractor loses it
SCOPE OF WORK

What maintenance covers

Audit of the current state

What runs, which versions, where backups live, who has access and what is already near failure.

Server setup

Web server, database, certificates, permissions and separate production and staging environments.

Mail server

Domain, mailboxes, SPF, DKIM and DMARC, antispam and reputation checks so mail arrives.

Backups

Schedule, retention, an off-server copy and regular restore tests.

Monitoring

Availability, disk, memory, load, certificate expiry and the key services themselves, not just ping.

Updates and security

Scheduled system and application updates, restricted access, protection against password guessing.

Response rules

Who owns an incident, how fast we respond and what counts as critical for your business.

Reporting

What happened over the month, which incidents occurred and what was done to prevent repeats.

TECHNOLOGY LANDSCAPE

What we work with

LinuxThe main environment: web servers, databases, queues and background jobs
Mail stackSMTP, IMAP, antispam, DKIM signing and a DMARC policy for the domain
TLSCertificates with automatic renewal and expiry tracking
Backups and storageDatabase and file backups stored away from the main server
Monitoring and alertsService checks, warning thresholds and alerts to the on-call engineer
LogsLog collection and rotation, so incidents are diagnosed from facts

The server, the domain and the hosting stay registered to your company. We receive access for the work, and you can revoke it at any point.

QUESTIONS

What people usually ask

Why does mail from our domain end up in spam?

Usually three reasons: the domain does not sign its mail (no DKIM), it does not declare who may send on its behalf (no SPF), or the server is on blocklists from past campaigns. We check all three, configure signing and policy, then watch the domain's reputation over time — a one-off fix does not hold.

We already pay for hosting. Why pay for maintenance too?

Hosting is responsible for the hardware being on and the network working. Everything above that — versions, configuration, backups, mail, updates and incident response — belongs to the server's owner. If nobody handles it, the problem surfaces when it is already in the way.

How fast do you respond at night?

That is agreed in writing before the work starts: what counts as critical for you and how fast we respond. Monitoring runs around the clock and critical checks page the on-call engineer immediately — most incidents close before your customers hear about them.

What if the server does go down?

A written procedure runs: we identify what exactly is unavailable, bring the service back or restore from a copy, and report status and recovery time to you. Backups have already been verified by restoring them, so the recovery time is known rather than discovered mid-outage.

Can the server be moved to another provider?

Yes, and it is routine work. We prepare the new environment, migrate data and configuration, verify everything in parallel and only then switch the domain. The move is scheduled for the quietest hours, and the old environment stays available until everything is confirmed working.

Related capabilities

Corporate email Domain, mailboxes, delivery WEB development What runs on the server API integrations Exchanges that must not fail Email campaigns Domain reputation

Let’s apply this capability to your project

We define the challenge, dependencies and a realistic delivery plan.

Discuss a project