<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Al’s-Ops LLC</title>
    <description>The latest articles on DEV Community by Al’s-Ops LLC (als-ops).</description>
    <link>https://dev.to/als-ops</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Forganization%2Fprofile_image%2F16314%2F34cff913-1704-490a-be98-01c93da5d12b.png</url>
      <title>DEV Community: Al’s-Ops LLC</title>
      <link>https://dev.to/als-ops</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/als-ops"/>
    <language>en</language>
    <item>
      <title>SSH hardening for Linux servers: settings you can verify</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:07:12 +0000</pubDate>
      <link>https://dev.to/als-ops/ssh-hardening-for-linux-servers-settings-you-can-verify-4c1e</link>
      <guid>https://dev.to/als-ops/ssh-hardening-for-linux-servers-settings-you-can-verify-4c1e</guid>
      <description>&lt;p&gt;Most SSH hardening guides are lists of settings to paste. The catch is that sshd quietly ignores a setting when an earlier one wins, so a pasted config can look right and change nothing.&lt;/p&gt;

&lt;p&gt;Keep an existing session open while you work and test every change from a second terminal.&lt;/p&gt;

&lt;h2&gt;
  
  
  sshd uses the first value it reads
&lt;/h2&gt;

&lt;p&gt;For most keywords, sshd uses the first value it reads, not the last (&lt;a href="https://man.openbsd.org/sshd_config" rel="noopener noreferrer"&gt;sshd_config(5)&lt;/a&gt;). On Debian and Ubuntu, &lt;code&gt;/etc/ssh/sshd_config&lt;/code&gt; starts with &lt;code&gt;Include /etc/ssh/sshd_config.d/*.conf&lt;/code&gt;, read in lexical order. A drop-in called &lt;code&gt;10-hardening.conf&lt;/code&gt; beats the main file and any drop-in that sorts after it.&lt;/p&gt;

&lt;p&gt;The reliable check is to ask sshd itself:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo sshd -t   &lt;span class="c"&gt;# syntax check, silent when valid&lt;/span&gt;
sudo sshd -T &lt;span class="o"&gt;|&lt;/span&gt; grep -iE &lt;span class="s2"&gt;"passwordauthentication|permitrootlogin"&lt;/span&gt;   &lt;span class="c"&gt;# effective values&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Baseline
&lt;/h2&gt;

&lt;p&gt;Create the login group first, or &lt;code&gt;AllowGroups&lt;/code&gt; locks you out:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo groupadd ssh-users
sudo usermod -aG ssh-users deploy
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;





&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight ssh"&gt;&lt;code&gt;&lt;span class="c1"&gt;# /etc/ssh/sshd_config.d/10-hardening.conf
&lt;/span&gt;&lt;span class="nc"&gt;PasswordAuthentication&lt;/span&gt; &lt;span class="n"&gt;no&lt;/span&gt;
&lt;span class="nc"&gt;KbdInteractiveAuthentication&lt;/span&gt; &lt;span class="n"&gt;no&lt;/span&gt;
&lt;span class="nc"&gt;PermitRootLogin&lt;/span&gt; &lt;span class="n"&gt;no&lt;/span&gt;
&lt;span class="nc"&gt;AllowGroups&lt;/span&gt; &lt;span class="n"&gt;ssh&lt;/span&gt;&lt;span class="o"&gt;-&lt;/span&gt;&lt;span class="n"&gt;users&lt;/span&gt;
&lt;span class="nc"&gt;MaxAuthTries&lt;/span&gt; &lt;span class="mi"&gt;3&lt;/span&gt;
&lt;span class="nc"&gt;LoginGraceTime&lt;/span&gt; &lt;span class="mi"&gt;30&lt;/span&gt;
&lt;span class="nc"&gt;X11Forwarding&lt;/span&gt; &lt;span class="n"&gt;no&lt;/span&gt;
&lt;span class="nc"&gt;AllowAgentForwarding&lt;/span&gt; &lt;span class="n"&gt;no&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;





&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo sshd -t &lt;span class="o"&gt;&amp;amp;&amp;amp;&lt;/span&gt; sudo systemctl reload ssh   &lt;span class="c"&gt;# sshd on RHEL/Fedora&lt;/span&gt;
ssh -o &lt;span class="nv"&gt;PubkeyAuthentication&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;no -o &lt;span class="nv"&gt;PreferredAuthentications&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;password deploy@your-server
&lt;span class="c"&gt;# expect: Permission denied (publickey)&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Before OpenSSH 10.6, every key your agent offers counts against &lt;code&gt;MaxAuthTries&lt;/code&gt;, so pin keys per host with &lt;code&gt;IdentityFile&lt;/code&gt; + &lt;code&gt;IdentitiesOnly yes&lt;/code&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Built-in brute-force penalties (OpenSSH 9.8+)
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://www.openssh.com/txt/release-9.8" rel="noopener noreferrer"&gt;OpenSSH 9.8&lt;/a&gt; added &lt;code&gt;PerSourcePenalties&lt;/code&gt;, on by default. On 9.8 or later you already have basic throttling; exempt a shared office/VPN range with &lt;code&gt;PerSourcePenaltyExemptList&lt;/code&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  Limit where SSH is reachable
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo ufw allow from 203.0.113.0/24 to any port 22 proto tcp
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Changing the port only cuts log noise. On Ubuntu 22.10+ SSH is socket-activated: run &lt;code&gt;systemctl daemon-reload&lt;/code&gt; then &lt;code&gt;systemctl restart ssh.socket&lt;/code&gt; (&lt;a href="https://bugs.launchpad.net/ubuntu/+source/openssh/+bug/2069041" rel="noopener noreferrer"&gt;LP#2069041&lt;/a&gt;).&lt;/p&gt;

&lt;h2&gt;
  
  
  Hardware keys, crypto defaults, patching
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;ssh-keygen -t ed25519-sk&lt;/code&gt; puts the key on a FIDO token (OpenSSH 8.2+).&lt;/li&gt;
&lt;li&gt;Don't paste old &lt;code&gt;Ciphers&lt;/code&gt;/&lt;code&gt;KexAlgorithms&lt;/code&gt; lists: they drop the default post-quantum hybrid &lt;code&gt;mlkem768x25519-sha256&lt;/code&gt;. Audit with &lt;a href="https://github.com/jtesta/ssh-audit" rel="noopener noreferrer"&gt;ssh-audit&lt;/a&gt; and remove items with a leading &lt;code&gt;-&lt;/code&gt;. On RHEL use &lt;code&gt;update-crypto-policies&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;a href="https://nvd.nist.gov/vuln/detail/CVE-2024-6387" rel="noopener noreferrer"&gt;CVE-2024-6387&lt;/a&gt; (regreSSHion) hit portable OpenSSH 8.5p1–9.7p1. Distros backport fixes, so check your distro's tracker, not the version string.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Watch successful logins
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo journalctl -u ssh --since yesterday &lt;span class="o"&gt;|&lt;/span&gt; grep Accepted
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Since 9.8 some lines are tagged &lt;code&gt;sshd-session&lt;/code&gt;; filtering by unit catches both.&lt;/p&gt;

&lt;p&gt;Full guide with the reasoning for each setting and a verification table: &lt;a href="https://alsopss.com/blog/ssh-hardening" rel="noopener noreferrer"&gt;https://alsopss.com/blog/ssh-hardening&lt;/a&gt;&lt;/p&gt;

</description>
      <category>linux</category>
      <category>security</category>
      <category>ssh</category>
      <category>devops</category>
    </item>
    <item>
      <title>Fail2ban alternatives for Linux servers</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:06:32 +0000</pubDate>
      <link>https://dev.to/als-ops/fail2ban-alternatives-for-linux-servers-2do3</link>
      <guid>https://dev.to/als-ops/fail2ban-alternatives-for-linux-servers-2do3</guid>
      <description>&lt;p&gt;Fail2ban reads log files, spots repeated failures such as bad SSH passwords, and adds a firewall rule to ban the address for a while. It is free, widely packaged and does that one job well. People look for alternatives for a few recurring reasons.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why people look elsewhere
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Each server bans on its own. An address banned on one server can keep attacking the others.&lt;/li&gt;
&lt;li&gt;It fails silently. If a log path or format changes, fail2ban keeps running and simply matches nothing, which looks the same as a quiet week.&lt;/li&gt;
&lt;li&gt;It only blocks. It doesn’t tell you about the login that succeeded, the process that started afterwards or the file that changed.&lt;/li&gt;
&lt;li&gt;Regex filters and jail files take time to tune and to keep working across distributions.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The alternatives
&lt;/h2&gt;

&lt;h3&gt;
  
  
  SSHGuard
&lt;/h3&gt;

&lt;p&gt;A small daemon written in C that watches logs for SSH and a few other services and blocks attackers through your firewall. It is lighter and simpler than fail2ban, with less to configure, but it has the same scope: one host, blocking only.&lt;/p&gt;

&lt;h3&gt;
  
  
  CrowdSec
&lt;/h3&gt;

&lt;p&gt;An open-source engine that parses logs with community-maintained scenarios and hands decisions to “bouncers” that enforce them in the firewall, a web server or a CDN. Its main difference is the community blocklist: addresses reported by other users can be blocked before they reach you. The engine is free; a hosted console adds paid features. &lt;a href="https://vigilhold.com/compare/crowdsec" rel="noopener noreferrer"&gt;Vigilhold vs CrowdSec, in detail&lt;/a&gt;&lt;/p&gt;

&lt;h3&gt;
  
  
  Wazuh active response
&lt;/h3&gt;

&lt;p&gt;Wazuh is a full open-source security platform: log analysis, file integrity monitoring, vulnerability detection and more, with “active response” scripts that can block addresses. It does far more than fail2ban, and asks far more of you: a manager, an indexer and a dashboard to install, size and keep updated. &lt;a href="https://vigilhold.com/compare/wazuh" rel="noopener noreferrer"&gt;Vigilhold vs Wazuh, in detail&lt;/a&gt;&lt;/p&gt;

&lt;h3&gt;
  
  
  Vigilhold
&lt;/h3&gt;

&lt;p&gt;Our hosted host agent for Linux servers. When a response playbook bans an address, the ban goes to every server in your account, and servers that were offline pick it up when they reconnect. It also watches logins, processes, connections and file changes, and warns you if a server with SSH open to the internet stops reporting login activity for three hours, which catches the silent-failure case above. &lt;a href="https://vigilhold.com/compare/fail2ban" rel="noopener noreferrer"&gt;Vigilhold vs fail2ban, in detail&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Side by side
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;&lt;/th&gt;
&lt;th&gt;fail2ban&lt;/th&gt;
&lt;th&gt;SSHGuard&lt;/th&gt;
&lt;th&gt;CrowdSec&lt;/th&gt;
&lt;th&gt;Wazuh&lt;/th&gt;
&lt;th&gt;Vigilhold&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Blocks attacking IPs&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;td&gt;Yes, via active response&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Bans shared across your servers&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;td&gt;Configurable&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Detects what happens after a login&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Warns when log monitoring goes quiet&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;No&lt;/td&gt;
&lt;td&gt;Not by default&lt;/td&gt;
&lt;td&gt;Not by default&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;You run the server side&lt;/td&gt;
&lt;td&gt;Nothing to run&lt;/td&gt;
&lt;td&gt;Nothing to run&lt;/td&gt;
&lt;td&gt;Optional&lt;/td&gt;
&lt;td&gt;Yes&lt;/td&gt;
&lt;td&gt;No, hosted&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Cost&lt;/td&gt;
&lt;td&gt;Free&lt;/td&gt;
&lt;td&gt;Free&lt;/td&gt;
&lt;td&gt;Free engine, paid console&lt;/td&gt;
&lt;td&gt;Free, or paid cloud&lt;/td&gt;
&lt;td&gt;From $39/month for 5 servers&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Which one to pick
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;One or two personal servers and fail2ban works: keep it, and check fail2ban-client status now and then to confirm it is still counting failures.&lt;/li&gt;
&lt;li&gt;You want fail2ban’s job with less to configure: SSHGuard.&lt;/li&gt;
&lt;li&gt;You want blocklists informed by many other servers: CrowdSec.&lt;/li&gt;
&lt;li&gt;You have the people to run a security platform and need broad coverage: Wazuh.&lt;/li&gt;
&lt;li&gt;A small team running a fleet that wants bans shared everywhere and detection beyond blocking, without running a server for it: Vigilhold, with a 14-day trial and no card.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Setting up a new server? The checklist covers fail2ban settings that commonly break on recent distributions. &lt;a href="https://alsopss.com/blog/secure-linux-vps-checklist" rel="noopener noreferrer"&gt;How to secure a Linux VPS&lt;/a&gt;&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;A ban tool that silently stops matching is worse than none, because it looks like protection. Whatever you choose, check that it is still counting.&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/fail2ban-alternatives" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>security</category>
      <category>linux</category>
      <category>guide</category>
    </item>
    <item>
      <title>How to secure a Linux VPS: a practical checklist</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:05:51 +0000</pubDate>
      <link>https://dev.to/als-ops/how-to-secure-a-linux-vps-a-practical-checklist-3ope</link>
      <guid>https://dev.to/als-ops/how-to-secure-a-linux-vps-a-practical-checklist-3ope</guid>
      <description>&lt;p&gt;A new server on a public IP address starts receiving SSH login attempts within minutes. Almost all of them are automated and almost all of them fail, as long as a few basics are in place. This checklist covers those basics for Debian and Ubuntu; the ideas are the same on other distributions.&lt;/p&gt;

&lt;p&gt;Keep your current SSH session open while you change SSH or firewall settings, and test from a second terminal. That way a mistake doesn’t lock you out.&lt;/p&gt;

&lt;h2&gt;
  
  
  1. Install updates, and keep installing them
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo apt update &lt;span class="o"&gt;&amp;amp;&amp;amp;&lt;/span&gt; sudo apt upgrade
sudo apt install unattended-upgrades
sudo dpkg-reconfigure -plow unattended-upgrades
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Automatic security updates close known holes without anyone remembering to log in. Reboot when a kernel update asks for it.&lt;/p&gt;

&lt;h2&gt;
  
  
  2. Log in as a normal user with an SSH key
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo adduser deploy
sudo usermod -aG sudo deploy
&lt;span class="c"&gt;# from your own computer:&lt;/span&gt;
ssh-copy-id deploy@your-server
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Check that you can log in as that user with the key before going further.&lt;/p&gt;

&lt;h2&gt;
  
  
  3. Turn off password and root logins
&lt;/h2&gt;

&lt;p&gt;Put your settings in a drop-in file so package updates don’t overwrite them:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="c"&gt;# /etc/ssh/sshd_config.d/10-hardening.conf&lt;/span&gt;
PasswordAuthentication no
KbdInteractiveAuthentication no
PermitRootLogin no
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;





&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo sshd -t &lt;span class="o"&gt;&amp;amp;&amp;amp;&lt;/span&gt; sudo systemctl reload ssh
sudo sshd -T &lt;span class="o"&gt;|&lt;/span&gt; grep -E &lt;span class="s2"&gt;"passwordauthentication|permitrootlogin"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The last command prints the settings sshd is really using. If it still says yes, another file is setting the value first: sshd uses the first value it reads, so check the other files in sshd_config.d. On RHEL-based systems the service is called sshd instead of ssh.&lt;/p&gt;

&lt;h2&gt;
  
  
  4. Close every port you don’t need
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo ss -tulpn
sudo apt install ufw
sudo ufw default deny incoming
sudo ufw default allow outgoing
sudo ufw allow OpenSSH
sudo ufw allow 443/tcp   &lt;span class="c"&gt;# only if this server serves HTTPS&lt;/span&gt;
sudo ufw &lt;span class="nb"&gt;enable&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;ss shows what is listening. Anything listening on 0.0.0.0 that only needs to be reached locally, such as a database, should listen on 127.0.0.1 instead.&lt;/p&gt;

&lt;p&gt;Running Docker? Ports published with -p are opened by Docker’s own firewall rules and are reachable even when ufw says they are blocked. Publish to 127.0.0.1 (for example -p 127.0.0.1:5432:5432) unless the port really should be public.&lt;/p&gt;

&lt;h2&gt;
  
  
  5. Block brute-force attempts
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="c"&gt;# /etc/fail2ban/jail.local&lt;/span&gt;
&lt;span class="o"&gt;[&lt;/span&gt;sshd&lt;span class="o"&gt;]&lt;/span&gt;
enabled &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="kc"&gt;true&lt;/span&gt;
backend &lt;span class="o"&gt;=&lt;/span&gt; systemd
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;





&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;sudo apt install fail2ban
sudo systemctl restart fail2ban
sudo fail2ban-client status sshd
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;backend = systemd matters on systems that only log to the journal, which is the default on recent Debian. Without it the sshd jail has no log file to read. The status command should show a non-zero “Total failed” count within a day on any public server; if it stays at zero, the jail isn’t seeing your logs. &lt;a href="https://alsopss.com/blog/fail2ban-alternatives" rel="noopener noreferrer"&gt;Alternatives to fail2ban&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  6. Keep backups somewhere else
&lt;/h2&gt;

&lt;p&gt;A backup stored on the same server disappears with it. Copy backups to another provider or account, and restore one now and then to prove it works.&lt;/p&gt;

&lt;h2&gt;
  
  
  7. Watch for what gets through
&lt;/h2&gt;

&lt;p&gt;Everything above lowers the odds of a break-in; none of it tells you if one happens. At minimum, look at recent logins and new listening ports from time to time:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;last -a &lt;span class="o"&gt;|&lt;/span&gt; head
sudo journalctl -u ssh --since &lt;span class="s2"&gt;"24 hours ago"&lt;/span&gt; &lt;span class="o"&gt;|&lt;/span&gt; grep Accepted
sudo ss -tulpn
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Past a handful of servers, that turns into a job nobody has time for. A host-based intrusion detection system does the watching for you. &lt;a href="https://alsopss.com/blog/host-based-intrusion-detection-system" rel="noopener noreferrer"&gt;What a HIDS does&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The short version
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Step&lt;/th&gt;
&lt;th&gt;Check it worked&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Automatic updates&lt;/td&gt;
&lt;td&gt;systemctl list-timers apt-daily-upgrade shows a next run&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Key-only SSH, no root&lt;/td&gt;
&lt;td&gt;sshd -T shows passwordauthentication no and permitrootlogin no&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Firewall&lt;/td&gt;
&lt;td&gt;ufw status lists only the ports you meant to open&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Brute-force blocking&lt;/td&gt;
&lt;td&gt;fail2ban-client status sshd shows failed attempts being counted&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Backups&lt;/td&gt;
&lt;td&gt;You restored one recently&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Monitoring&lt;/td&gt;
&lt;td&gt;You would know about a new login today, not next month&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Vigilhold, our host agent for Linux servers, covers step 7 across a whole fleet: it watches logins, processes, connections and file changes on every server, and warns you if a server with SSH open stops reporting login activity, which is how a broken setup like a blind fail2ban jail usually shows up. &lt;a href="https://alsopss.com/products" rel="noopener noreferrer"&gt;See Vigilhold&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/secure-linux-vps-checklist" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>linux</category>
      <category>security</category>
      <category>devops</category>
      <category>selfhosted</category>
    </item>
    <item>
      <title>What is a host-based intrusion detection system (HIDS)?</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:05:10 +0000</pubDate>
      <link>https://dev.to/als-ops/what-is-a-host-based-intrusion-detection-system-hids-1g8</link>
      <guid>https://dev.to/als-ops/what-is-a-host-based-intrusion-detection-system-hids-1g8</guid>
      <description>&lt;p&gt;A host-based intrusion detection system (HIDS) is software that runs on the server itself and watches for signs that someone has broken in or is trying to. Where a network intrusion detection system watches traffic passing a point on the network, a HIDS sees what actually happens on the machine: who logged in, what ran, what connected out and which files changed.&lt;/p&gt;

&lt;h2&gt;
  
  
  What a HIDS watches
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Logins: successful and failed SSH logins, new users, sudo use, logins from places or at times that don’t fit.&lt;/li&gt;
&lt;li&gt;Processes: new or unusual programs, especially ones started by a web server or running from /tmp.&lt;/li&gt;
&lt;li&gt;Network connections: listening ports that appear, and outbound connections to addresses that have no business being contacted.&lt;/li&gt;
&lt;li&gt;File integrity: changes to system binaries, SSH keys, cron jobs, systemd units and other places attackers use to stay in.&lt;/li&gt;
&lt;li&gt;Logs: authentication, kernel and application logs, matched against known attack patterns.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Most real intrusions on Linux servers leave traces in several of these at once: a brute-forced login, then a download, then a new cron entry or SSH key so the attacker can come back. A HIDS is useful because it can connect those steps on the host where they happen.&lt;/p&gt;

&lt;h2&gt;
  
  
  HIDS vs NIDS
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;&lt;/th&gt;
&lt;th&gt;Host-based (HIDS)&lt;/th&gt;
&lt;th&gt;Network-based (NIDS)&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Runs on&lt;/td&gt;
&lt;td&gt;Each server&lt;/td&gt;
&lt;td&gt;A network tap, span port or gateway&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Sees&lt;/td&gt;
&lt;td&gt;Logins, processes, files, local logs&lt;/td&gt;
&lt;td&gt;Packets and flows between machines&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Encrypted traffic&lt;/td&gt;
&lt;td&gt;Sees what happens after it is decrypted on the host&lt;/td&gt;
&lt;td&gt;Mostly sees metadata only&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Blind spot&lt;/td&gt;
&lt;td&gt;Hosts without an agent&lt;/td&gt;
&lt;td&gt;What happens inside a host&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Fits&lt;/td&gt;
&lt;td&gt;Cloud and VPS servers you don’t control the network for&lt;/td&gt;
&lt;td&gt;Networks you own end to end&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;For servers rented from a cloud or VPS provider you usually can’t put a sensor on the network at all, so host-based detection is the practical option.&lt;/p&gt;

&lt;h2&gt;
  
  
  HIDS vs EDR vs SIEM
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;HIDS: detects and alerts on activity on one host. Classic examples are OSSEC and file integrity checkers like AIDE.&lt;/li&gt;
&lt;li&gt;EDR (endpoint detection and response): a HIDS plus the ability to respond, such as killing a process or isolating a machine, usually with a central console. Many EDR products were built for laptops and Windows first.&lt;/li&gt;
&lt;li&gt;SIEM: collects logs from everything into one place to search and correlate. It depends on agents or log shippers on the hosts to feed it, and someone has to read what it finds.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Options for Linux servers
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Tool&lt;/th&gt;
&lt;th&gt;What it is&lt;/th&gt;
&lt;th&gt;Good to know&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;OSSEC&lt;/td&gt;
&lt;td&gt;The original open-source HIDS: log analysis, file integrity, rootkit checks&lt;/td&gt;
&lt;td&gt;Mature and still actively released&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Wazuh&lt;/td&gt;
&lt;td&gt;Open-source platform that grew out of OSSEC, with agents, a manager, an indexer and a dashboard&lt;/td&gt;
&lt;td&gt;Very capable; you run and maintain the server side&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;AIDE&lt;/td&gt;
&lt;td&gt;File integrity checker that compares files against a stored database&lt;/td&gt;
&lt;td&gt;Detects changes only; needs scheduling and someone to read reports&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;auditd&lt;/td&gt;
&lt;td&gt;The Linux kernel audit framework&lt;/td&gt;
&lt;td&gt;Records almost anything; turning records into alerts is up to you&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Falco&lt;/td&gt;
&lt;td&gt;Runtime detection from system calls, aimed at containers and Kubernetes&lt;/td&gt;
&lt;td&gt;Rules engine; alerts need routing somewhere&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;osquery&lt;/td&gt;
&lt;td&gt;Exposes system state as SQL tables&lt;/td&gt;
&lt;td&gt;Great for asking questions; not an alerting product on its own&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Vigilhold&lt;/td&gt;
&lt;td&gt;Hosted agent and console from Al’s-Ops&lt;/td&gt;
&lt;td&gt;Detection plus limited response you approve; from $39/month for five servers&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Tools like fail2ban, SSHGuard and CrowdSec are often mentioned in the same breath. They block addresses that attack a service, which is prevention for one kind of attack rather than intrusion detection, but they pair well with a HIDS. &lt;a href="https://alsopss.com/blog/fail2ban-alternatives" rel="noopener noreferrer"&gt;Compare fail2ban alternatives&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What to look for
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Alerts a person will read: grouped repeats and plain explanations, not one message per log line.&lt;/li&gt;
&lt;li&gt;Coverage you can check: a way to see that every server is reporting, and a warning when one goes quiet.&lt;/li&gt;
&lt;li&gt;Response you control: automatic actions limited to safe ones, with anything disruptive waiting for approval.&lt;/li&gt;
&lt;li&gt;Setup time that matches your team: an open-source stack is free to download but someone has to run it.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Where Vigilhold fits
&lt;/h2&gt;

&lt;p&gt;Vigilhold is a host agent for Linux servers, including Docker hosts, with one web console for all of them. It watches logins, processes, connections and file changes, explains critical and warning alerts in two sentences, and asks before anything disruptive. When a response playbook bans an address, the ban reaches every server in the account. It installs with one command per host. &lt;a href="https://alsopss.com/products" rel="noopener noreferrer"&gt;How Vigilhold works&lt;/a&gt;&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;A HIDS answers one question a firewall can’t: what actually happened on this server?&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/host-based-intrusion-detection-system" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>security</category>
      <category>linux</category>
      <category>guide</category>
    </item>
    <item>
      <title>Domain trust alone is not enough</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:04:29 +0000</pubDate>
      <link>https://dev.to/als-ops/domain-trust-alone-is-not-enough-55c</link>
      <guid>https://dev.to/als-ops/domain-trust-alone-is-not-enough-55c</guid>
      <description>&lt;p&gt;HTTPS answers whether a certificate is authorized for a domain name. The tempting assumption is that this is enough to trust the machine on the other end of the connection.&lt;/p&gt;

&lt;p&gt;It fails when DNS is hijacked, a registrar path is abused, or validation still succeeds while traffic terminates on an attacker-controlled host. The client can see a valid domain certificate and still not be talking to an authorized server.&lt;/p&gt;

&lt;p&gt;We wrote up Double Sign (Double Certificate Authority): require a second, independent cryptographic authorization for the node — not only the domain — before the client accepts the connection.&lt;/p&gt;

&lt;p&gt;AL'S-OPS LLC is a Utah company building privacy-first infrastructure software.&lt;/p&gt;

&lt;h2&gt;
  
  
  What changed
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Domain authentication stays on the usual Web PKI path (TLS certificate for the name).&lt;/li&gt;
&lt;li&gt;Node authentication is a separate credential binding a server key to an authorized IP, service, or node identity, validated against a trust anchor that is not the domain’s DNS or registrar path.&lt;/li&gt;
&lt;li&gt;The server must prove possession of the private key.&lt;/li&gt;
&lt;li&gt;Both paths must succeed. Domain-valid without node-valid is a reject. Node-valid does not authorize an unrelated domain.&lt;/li&gt;
&lt;li&gt;Trust anchor design separates a long-lived offline root from delegated operational keys, with signed epoch metadata, revocation, and anti-rollback patterns in the spirit of systems like The Update Framework (TUF).&lt;/li&gt;
&lt;li&gt;Baseline operational policy uses monthly trust epochs for credential renewal (and preferably key rotation), while compromised keys remain revocable immediately.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The real problem
&lt;/h2&gt;

&lt;p&gt;Proving control of a domain is not the same as proving authorization to operate the server that claims to represent that domain. A single validation path tied to DNS and registrar infrastructure is a single place that path can fail open for the attacker.&lt;/p&gt;

&lt;p&gt;Double Sign is meant to force a second defeat: the attacker needs a valid node credential and private key, not only a domain certificate that looks legitimate.&lt;/p&gt;

&lt;h2&gt;
  
  
  What we are not doing
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Replacing TLS, Certificate Transparency, RPKI, or DANE — those solve different parts of the problem.&lt;/li&gt;
&lt;li&gt;Claiming browser-wide deployment or a shipped product feature in this note.&lt;/li&gt;
&lt;li&gt;Treating a second signature as meaningful if it depends on the same administrators, DNS path, or trust root as the first.&lt;/li&gt;
&lt;li&gt;Inventing customers, fleet size, or adoption metrics.&lt;/li&gt;
&lt;/ul&gt;

&lt;blockquote&gt;
&lt;p&gt;Domain identity and server authorization should not share one quiet point of failure. If the second signature is not independent, it is not a second signature.&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/domain-trust-alone-is-not-enough" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>security</category>
      <category>tls</category>
      <category>pki</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title>The default is the product</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:03:48 +0000</pubDate>
      <link>https://dev.to/als-ops/the-default-is-the-product-3kfi</link>
      <guid>https://dev.to/als-ops/the-default-is-the-product-3kfi</guid>
      <description>&lt;p&gt;Most tools ship the active path on.&lt;/p&gt;

&lt;p&gt;Response is enabled. Automation is eager. The operator is expected to turn risk down after install. That is the easy default. It also means the first hour on a host is the one least reviewed and the most capable of doing harm.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;We do the inverse.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;AL'S-OPS LLC is a Utah company building privacy-first infrastructure software.&lt;/p&gt;

&lt;h2&gt;
  
  
  What changed
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Observe is the default on a new host.&lt;/li&gt;
&lt;li&gt;Response is a mode, not a surprise.&lt;/li&gt;
&lt;li&gt;If the mode is off, the system watches and records. It does not ban, kill, or revert.&lt;/li&gt;
&lt;li&gt;Destructive actions stay behind an explicit switch the operator can see.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  The real problem
&lt;/h2&gt;

&lt;p&gt;A silent default that acts is harder to audit than a silent default that does not.&lt;/p&gt;

&lt;p&gt;Install docs rarely get read in full. Dashboards get opened later. The path that runs before anyone looks is the product, whether you named it that way or not. If that path can change the host, the company owns the outcome.&lt;/p&gt;

&lt;h2&gt;
  
  
  What we are not doing
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Not shipping response on "for convenience"&lt;/li&gt;
&lt;li&gt;Not treating automation theater as a feature list&lt;/li&gt;
&lt;li&gt;Not asking operators to discover the off switch under pressure&lt;/li&gt;
&lt;li&gt;Not publishing counts from our own machines as proof that the default is safe&lt;/li&gt;
&lt;/ul&gt;

&lt;blockquote&gt;
&lt;p&gt;The default is the product. If you would not defend that path in the first hour, do not ship it on.&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/the-default-is-the-product" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>infrastructure</category>
      <category>engineering</category>
      <category>security</category>
      <category>reliability</category>
    </item>
    <item>
      <title>Dual-Write With the Flag Off</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:03:08 +0000</pubDate>
      <link>https://dev.to/als-ops/dual-write-with-the-flag-off-41bf</link>
      <guid>https://dev.to/als-ops/dual-write-with-the-flag-off-41bf</guid>
      <description>&lt;p&gt;The tempting version of a storage change is a cutover.&lt;/p&gt;

&lt;p&gt;Turn off the old path.&lt;/p&gt;

&lt;p&gt;Turn on the new path.&lt;/p&gt;

&lt;p&gt;Hope both still mean the same thing.&lt;/p&gt;

&lt;p&gt;That is how you get a clean demo and a dirty night.&lt;/p&gt;

&lt;p&gt;AL'S-OPS LLC is a Utah company building privacy-first infrastructure software. This week we did not cut over. We added a dual-write on the servers table and left the production flag off.&lt;/p&gt;

&lt;h2&gt;
  
  
  What shipped
&lt;/h2&gt;

&lt;p&gt;Two write paths — upsertServer() and markServerOffline() — also write to an in-memory mirror. The mirror sits behind MIRROR_DB_ENABLED, which defaults to off.&lt;/p&gt;

&lt;p&gt;Around that:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;A parity checker comparing Postgres to the mirror&lt;/li&gt;
&lt;li&gt;A queue-flush job&lt;/li&gt;
&lt;li&gt;A parity-check job&lt;/li&gt;
&lt;li&gt;A burn-in loop instead of assuming the first write was correct&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The new store is not live in production. That is intentional.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why the flag mattered
&lt;/h2&gt;

&lt;p&gt;A real bug showed up in the write-path SQL.&lt;/p&gt;

&lt;p&gt;Partial-column updates would have silently corrupted mirror data.&lt;/p&gt;

&lt;p&gt;Not a crash. Not a useful stack trace. Wrong state later, when you no longer remember which write did it.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;That is the class of failure dual-write is for. You want the new path to fail while the old path is still the source of truth.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;h2&gt;
  
  
  The rest of the same stretch
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;A stability and security sweep (9 fixes)&lt;/li&gt;
&lt;li&gt;Retirement of the previous store&lt;/li&gt;
&lt;li&gt;Early work on the next mirror layers — still dark&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Order mattered more than any one item: retire what you can, put the new path behind a flag, write to both, compare, then argue about turning it on.&lt;/p&gt;

&lt;h2&gt;
  
  
  The rule
&lt;/h2&gt;

&lt;p&gt;Storage migrations die in the gap between "it compiled" and "every write shape was tested."&lt;/p&gt;

&lt;p&gt;Partial updates look harmless. They are how two systems drift until neither copy is trusted.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Find that while the mirror is dark. Keep the flag off until parity is boring.&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/dual-write-with-the-flag-off" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>infrastructure</category>
      <category>database</category>
      <category>reliability</category>
      <category>engineering</category>
    </item>
    <item>
      <title>If the Safe Path Is Optional, It Will Not Get Used</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:02:27 +0000</pubDate>
      <link>https://dev.to/als-ops/if-the-safe-path-is-optional-it-will-not-get-used-31jl</link>
      <guid>https://dev.to/als-ops/if-the-safe-path-is-optional-it-will-not-get-used-31jl</guid>
      <description>&lt;p&gt;Most teams do not fail because they forgot a setting.&lt;/p&gt;

&lt;p&gt;They fail because the setting existed, the docs mentioned it, and the person who needed it was already busy.&lt;/p&gt;

&lt;p&gt;AL'S-OPS LLC is a Utah company building privacy-first security and infrastructure tools. One rule keeps showing up in the work: if the safe path is optional, it will not get used.&lt;/p&gt;

&lt;p&gt;Not because people are careless. Because optional safety is extra work, and extra work loses to whatever is already on.&lt;/p&gt;

&lt;h2&gt;
  
  
  Configuration is not the same as protection
&lt;/h2&gt;

&lt;p&gt;A product can look complete and still leave the hard choice to the user.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Turn on the safer mode.&lt;/li&gt;
&lt;li&gt;Tighten the default.&lt;/li&gt;
&lt;li&gt;Read the checklist.&lt;/li&gt;
&lt;li&gt;Do not click the dangerous thing.&lt;/li&gt;
&lt;/ul&gt;

&lt;blockquote&gt;
&lt;p&gt;That is not a design. That is a transfer of responsibility.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;The tired operator will not complete your configuration journey. They will accept the default and move on. If that default is loose, the product was loose. The checkbox you added later does not change that.&lt;/p&gt;

&lt;h2&gt;
  
  
  What this looks like in practice
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;A dashboard that flashes 0 while data is still loading. That reads as "nothing is wrong" when the truth is "we do not know yet." Empty or a loading state is the safer default.&lt;/li&gt;
&lt;li&gt;A first-run install that leaves a destructive action one click away, with the confirm buried in an advanced tab. Most people never open that tab.&lt;/li&gt;
&lt;li&gt;A privacy control that is off until someone finds it. If extra collection is the out-of-box state, the product chose that — not the user.&lt;/li&gt;
&lt;li&gt;A security tool that says "you can lock this down in settings." If the loose mode is what ships, the loose mode is what runs at 2 a.m.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Same pattern every time. The optional safer path loses.&lt;/p&gt;

&lt;h2&gt;
  
  
  What a strong default actually means
&lt;/h2&gt;

&lt;p&gt;A strong default is not "locked down until nobody can work."&lt;/p&gt;

&lt;p&gt;It is the path that is reasonable before anyone has time to think.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;The interface should not invent a zero while data is still loading.&lt;/li&gt;
&lt;li&gt;The first-run posture should not require a specialist.&lt;/li&gt;
&lt;li&gt;Privacy should not be a later toggle.&lt;/li&gt;
&lt;li&gt;Destructive actions should be obvious and reversible.&lt;/li&gt;
&lt;li&gt;Quiet failure should be rarer than loud honesty.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If a user has to opt into those things, the product is asking them to finish the engineering.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why this is uncomfortable to ship
&lt;/h2&gt;

&lt;p&gt;Weak defaults demo better.&lt;/p&gt;

&lt;p&gt;They look flexible.&lt;/p&gt;

&lt;p&gt;They look customizable.&lt;/p&gt;

&lt;p&gt;They let a conversation say "you can configure that."&lt;/p&gt;

&lt;p&gt;Strong defaults look opinionated. They also survive the first bad night.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;We would rather explain why something is on than explain why something bad happened because it was off.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;h2&gt;
  
  
  The short version
&lt;/h2&gt;

&lt;p&gt;Do not confuse a setting with a safeguard.&lt;/p&gt;

&lt;p&gt;If the safe path is buried, it is not a path. It is a footnote. Design the first state as if nobody will ever open the advanced tab. Most people will not.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;That is not pessimism. That is how software actually gets used.&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/if-the-safe-path-is-optional" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>company</category>
      <category>productdesign</category>
      <category>defaults</category>
      <category>security</category>
    </item>
    <item>
      <title>Design for the Tired Operator, Not the Impressed Buyer</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:01:46 +0000</pubDate>
      <link>https://dev.to/als-ops/design-for-the-tired-operator-not-the-impressed-buyer-4pgk</link>
      <guid>https://dev.to/als-ops/design-for-the-tired-operator-not-the-impressed-buyer-4pgk</guid>
      <description>&lt;p&gt;Most infrastructure software is designed for the room where the purchase happens.&lt;/p&gt;

&lt;p&gt;The lights are good.&lt;/p&gt;

&lt;p&gt;The laptop is charged.&lt;/p&gt;

&lt;p&gt;Nobody is getting paged.&lt;/p&gt;

&lt;p&gt;The person on the call wants to be convinced.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;That room is not where the product will live.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;AL'S-OPS LLC is a Utah company building privacy-first security and infrastructure tools. The person we design for is usually tired. They are on a phone. They are missing context. They do not want a story. They want to know whether the system is telling the truth.&lt;/p&gt;

&lt;h2&gt;
  
  
  Buyers and operators are not the same audience
&lt;/h2&gt;

&lt;p&gt;A buyer can tolerate a dense dashboard.&lt;/p&gt;

&lt;p&gt;An operator cannot tolerate a dashboard that flashes a zero while data is still loading.&lt;/p&gt;

&lt;p&gt;A buyer likes a long feature list.&lt;/p&gt;

&lt;p&gt;An operator likes one sentence that still makes sense at 2 a.m.&lt;/p&gt;

&lt;p&gt;A buyer asks what else the platform can do.&lt;/p&gt;

&lt;p&gt;An operator asks what happens when the control plane is late, the network is ugly, or a human clicks the wrong thing.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;If you only design for the first person, you will ship something that looks complete and behaves badly the first night it matters.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;h2&gt;
  
  
  What changes when you design for fatigue
&lt;/h2&gt;

&lt;p&gt;A few things become non-negotiable.&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Honesty in the interface — if the data is not loaded, do not pretend it is. Empty is better than false.&lt;/li&gt;
&lt;li&gt;Defaults that are safe — the tired person should not have to configure their way into a reasonable posture. The safe path should already be on.&lt;/li&gt;
&lt;li&gt;Language that survives stress — if a status, alert, or action needs a paragraph of explanation, it will get misread. Short beats clever.&lt;/li&gt;
&lt;li&gt;Control that stays with the human — autonomy is useful. Surprise is not. The operator should still be able to see what happened, reverse it, and understand why.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;None of that photographs well in a launch post. All of it matters more than the launch post.&lt;/p&gt;

&lt;h2&gt;
  
  
  The test we use
&lt;/h2&gt;

&lt;blockquote&gt;
&lt;p&gt;Would this still be usable if the person in front of it had slept four hours, was on mobile, and did not trust the last tool they used?&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;If the answer is no, the work is not finished. It is just dressed up.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Impressive demos are easy. Software that stays trustworthy when the operator is tired is the actual product.&lt;/p&gt;
&lt;/blockquote&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/design-for-the-tired-operator" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>company</category>
      <category>productdesign</category>
      <category>operators</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title>Who is AL'S-OPS?</title>
      <dc:creator>ALSOPS</dc:creator>
      <pubDate>Fri, 20 Nov 2026 02:01:06 +0000</pubDate>
      <link>https://dev.to/als-ops/who-is-als-ops-33j0</link>
      <guid>https://dev.to/als-ops/who-is-als-ops-33j0</guid>
      <description>&lt;p&gt;  &lt;iframe src="https://www.youtube.com/embed/7HJn8cNfXkc" width="710" height="399"&gt;
  &lt;/iframe&gt;
&lt;/p&gt;

&lt;p&gt;AL'S-OPS LLC is a software company focused on privacy-first security and infrastructure tools.&lt;/p&gt;

&lt;p&gt;We're not trying to be the loudest in the market. We're trying to build systems people can actually trust.&lt;/p&gt;

&lt;p&gt;A lot of software looks impressive in demos and falls apart in real conditions. In security and infrastructure especially, that gap matters. If a system is hard to understand, unpredictable under pressure, or weak around sensitive data, it creates more risk than it removes.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;That's the problem AL'S-OPS is built around.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;h2&gt;
  
  
  What we care about
&lt;/h2&gt;

&lt;p&gt;We keep the standards simple:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Clear behavior — a system should be understandable without a sales call&lt;/li&gt;
&lt;li&gt;Strong defaults — the safe path should be the default path&lt;/li&gt;
&lt;li&gt;Operator control — people running the system stay in charge&lt;/li&gt;
&lt;li&gt;Privacy as a foundation — not a marketing line added later&lt;/li&gt;
&lt;/ul&gt;

&lt;blockquote&gt;
&lt;p&gt;The goal is software that stays useful when conditions are messy, not just when everything is right.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;h2&gt;
  
  
  What we work on
&lt;/h2&gt;

&lt;p&gt;Today that work includes:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Host defense systems&lt;/li&gt;
&lt;li&gt;Secure data tooling&lt;/li&gt;
&lt;li&gt;Infrastructure projects designed for real operators&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;We keep the focus practical:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;What does the system do?&lt;/li&gt;
&lt;li&gt;Where does it run?&lt;/li&gt;
&lt;li&gt;Is it trustworthy under pressure?&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Where we're based
&lt;/h2&gt;

&lt;p&gt;AL'S-OPS is based in Utah.&lt;/p&gt;

&lt;p&gt;If you're interested in privacy-first security and infrastructure software built around being serious instead of around hype, you're in the right place.&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on the &lt;a href="https://alsopss.com/blog/who-is-als-ops" rel="noopener noreferrer"&gt;Al's-Ops blog&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>company</category>
      <category>security</category>
      <category>privacy</category>
    </item>
  </channel>
</rss>
